id stringlengths 36 36 | tier stringclasses 1
value | github_login stringclasses 179
values | created_at stringlengths 32 32 | status stringclasses 2
values | sha256 stringlengths 64 64 | source_bytes int64 341 260k | leaderboard_rank int64 | source large_stringlengths 341 260k | metadata_json large_stringlengths 645 734 | result_json large_stringlengths 5 4.8k |
|---|---|---|---|---|---|---|---|---|---|---|
04274bfa-93e3-4dbb-8d61-d444b37655e6 | easy | sapient-sapiens | 2026-08-19 19:26:17.710153+00:00 | succeeded | daa4fc1f1821b95b049e62d1602f10121592d66b9efbbb3e78e141bca4403b2b | 33,252 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "04274bfa-93e3-4dbb-8d61-d444b37655e6",
"created_at": "2026-08-19 19:26:17.710153+00:00",
"db_md5": "2553f964fd1803486430c2690edf5c49",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "1380f1ad-ec18-4eac-81f7-8446678ca1e7",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.139393473954093,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0053124999999999995
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.149887540831041,
"example_count": 120... |
0428e0ac-57aa-4050-b301-4bbbd03a355f | easy | erdavis0 | 2026-08-28 14:21:43.576944+00:00 | succeeded | eadd19360be6ecdb82c0d25f4d3447d162d2edd2acd789ba5e873fafaac8fc3b | 19,602 | null | """Four-code N-conditioned circulant operator over an X-initialized tape.
Public delimiters only route opaque categories into independent right-aligned
N, X, and T tapes. N makes a straight-through hard choice among four globally
learned cyclic operators; X alone initializes recurrent state; T is consulted
only after... | {
"id": "0428e0ac-57aa-4050-b301-4bbbd03a355f",
"created_at": "2026-08-28 14:21:43.576944+00:00",
"db_md5": "be2bf3989baac5e04978d28d9e6dbbb4",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "db60595a-c809-48d4-8196-806b0723afc5",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.016973528529265,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010000000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.960889091406168,
"example_count": 600,... |
042bb029-f979-42b8-bf29-4f63b3d4cc46 | easy | RevellFTW | 2026-08-26 17:08:07.944959+00:00 | failed | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "042bb029-f979-42b8-bf29-4f63b3d4cc46",
"created_at": "2026-08-26 17:08:07.944959+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "brtlnkrisztian",
"github_login": "RevellFTW",
"run_id": "c89b0b0c-c6e6-473f-8a3a-4b1bed7e3dd3",
"tier": "easy",
"dataset_id": "e1",
"status": "f... | null
|
042cba84-3ec7-40e8-ae08-855d3a17ac62 | easy | DDanlov | 2026-08-26 18:29:55.491830+00:00 | failed | ecdb91030251bdd1002a533c4b81f348c7f90496852ba09bf948e2496ad60df1 | 29,497 | null | import math
import time
from typing import Any
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ImportError:
Submission, OptimizerBundle, ModelSpec, OptimizerSpec... | {
"id": "042cba84-3ec7-40e8-ae08-855d3a17ac62",
"created_at": "2026-08-26 18:29:55.491830+00:00",
"db_md5": "cdbebdcfbf88d9c76494727ba6f7fe4d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c97b0880-25b3-412f-9a06-10c9438364e5",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
042e7012-d271-4ca5-8823-93aa221f2ed3 | easy | erdavis0 | 2026-08-26 03:46:51.792847+00:00 | succeeded | 6fe6fdf46bcdc9c76ffef64fbdc6b302f2262f8879ac4d7daea4e62dccf8b390 | 30,354 | null | """Latent-role continuous REV C5: hybrid token CE and sequence mixture.
Public delimiter syntax only routes opaque token categories into distinct,
right-aligned N, X, and T tapes. Learned local cells compile separate
interaction, carry, and modulus-conditioned workspaces. N is isolated from
interaction and carry and... | {
"id": "042e7012-d271-4ca5-8823-93aa221f2ed3",
"created_at": "2026-08-26 03:46:51.792847+00:00",
"db_md5": "14ff631f83062f36483f24b1bd94c4c9",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "c022f38c-cd4c-42dd-a3ac-bfe4e852c9ad",
"tier": "easy",
"dataset_id": "e9",
"status": "succeeded",... | {
"score": {
"mean_loss": 6.905858993530273,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06666667014360428
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.053086280822754,
"example_count": 90,
... |
042f50df-81ed-4424-ad43-716524be9666 | easy | DDanlov | 2026-08-15 13:57:02.790551+00:00 | succeeded | 1e394ce0e09a975e4a9792c3a43683a6b525bab4da2a041387f3e282465bf78f | 18,847 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Pure No-Norm, No-Skip DKS Recurrent Core with Input RMS Normalization,
Micro-Step Recurrence (k_micro * T), Power-of-10 Place-Value Decoder, and RohanShampoo.
"""
from __future__ import annotations
import math
from typing import Optional, Tuple
impor... | {
"id": "042f50df-81ed-4424-ad43-716524be9666",
"created_at": "2026-08-15 13:57:02.790551+00:00",
"db_md5": "de0753598b89cba45f2fb9241be6780c",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "77c2694f-d33d-45e8-8509-f7910fc9a9f6",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 11.45116158403523,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016666666865348817
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.056649464279857,
"example_count": 100... |
044287bf-4bd9-43e7-9de7-70a7abbb0da5 | easy | shreyash-chonkie | 2026-09-01 02:43:34.385760+00:00 | succeeded | 5e77672079ea4fdf01900407ea9b522a1640462e7d16284c9f9ae1f600729bcb | 21,414 | null | """Learned token scanner feeding a random recurrent path reservoir."""
from __future__ import annotations
from itertools import permutations
import torch
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_s... | {
"id": "044287bf-4bd9-43e7-9de7-70a7abbb0da5",
"created_at": "2026-09-01 02:43:34.385760+00:00",
"db_md5": "c3437a3a0f44713f017e29f9299db178",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "be777375-e055-44e3-90ab-67b874173745",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.7637560367584229,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02666666731238365
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8119738101959229,
"example_count": 100... |
044328db-5455-44ad-ac8f-bea78c394df8 | easy | ddoan | 2026-08-14 10:59:43.117037+00:00 | succeeded | 419f020949fa64a81c8d3ef030ca82abf645279a73f344ef99b66721d7d4193c | 10,224 | null | """Place-aligned active memory with hard-gated diagonal CGRU updates."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_sta... | {
"id": "044328db-5455-44ad-ac8f-bea78c394df8",
"created_at": "2026-08-14 10:59:43.117037+00:00",
"db_md5": "4ed4c6d3cccff689f11a89c75d54fe61",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "13c76c22-5e2a-4d82-9e0c-97a229781997",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.1438318905686264,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1505447769044674,
"example_count": 80... |
0444b19b-4895-46e9-adf2-43cc9a9d1d71 | easy | nikhilkumar26 | 2026-08-29 18:07:43.794594+00:00 | succeeded | 5fc78342f6671b3bd62075bc0b7de606d19801a98f2636d1b7563960258d93c1 | 7,684 | null | """
One Layer Deeper Easy E1:
digit-wise learned square F (the 7.7% exact model) plus batch reuse.
x and N stay 8 digit slots. Shared F is one learned transition.
T chooses how many times F is applied. Batch reuse buys more
AdamW steps on the only architecture that generalized.
No modular arithmetic in Python.
"""
fr... | {
"id": "0444b19b-4895-46e9-adf2-43cc9a9d1d71",
"created_at": "2026-08-29 18:07:43.794594+00:00",
"db_md5": "3da0b62604617cbef046142d00eff798",
"submitter": "Nikhil Kumar",
"github_login": "nikhilkumar26",
"run_id": "50bb887c-7c49-41e6-8ac5-4a3873e33565",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 5.490152835845947,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02166666742414236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.6708898544311523,
"example_count": 100,... |
0448c88b-09a3-4dd3-badb-bc12991ea7a2 | easy | ehonig | 2026-08-20 09:20:01.520359+00:00 | succeeded | 55489eeee15489043622955ac6c57744dd85a6d2d917e80f747815dfe1bc841d | 23,592 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "0448c88b-09a3-4dd3-badb-bc12991ea7a2",
"created_at": "2026-08-20 09:20:01.520359+00:00",
"db_md5": "1a248aa4f5369903aba9a162c10b7b41",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "e3a92fc0-8f37-40d3-a134-6d63ce65f0c1",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.108075737953186,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07000000029802322
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1168811321258545,
"example_count": 100,... |
045c1cca-ea51-4faf-a6a6-b06b9e7bda41 | easy | 0Chris5R | 2026-08-07 12:45:53.380841+00:00 | succeeded | 27d4136cd2086ded14b8f521711d1861e08b4df19229bedab38cd1b9e65ebd92 | 20,838 | null | """Shared-digit Neural GPU with a compute-matched operator curriculum.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the ... | {
"id": "045c1cca-ea51-4faf-a6a6-b06b9e7bda41",
"created_at": "2026-08-07 12:45:53.380841+00:00",
"db_md5": "ff8617dabf6c97867b4c258bd136f00f",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "ac360b83-1604-4f63-9d2a-69ba3f092789",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.9442251789340435,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0416666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.865245258861593,
"example_count": 600,
... |
045e64af-e9b6-46b7-ac90-d6dff4ed549d | easy | jordanrubin | 2026-08-05 18:38:17.449926+00:00 | succeeded | 2214473aabeee3e5a3252928aedd50993436405bf900fc579ecc4ef88811f4f4 | 14,788 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "045e64af-e9b6-46b7-ac90-d6dff4ed549d",
"created_at": "2026-08-05 18:38:17.449926+00:00",
"db_md5": "f917ba4387feee357d1f6822e4a7825f",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "576ba648-56c7-45d5-b35e-f2ce0ea4fa67",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.150228049217347,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00458333323088785
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.152570554360146,
"example_count": 600,
... |
046719e6-a33a-4933-9ba3-b44f553b126d | easy | KaustubhKumar05 | 2026-08-25 16:36:48.166946+00:00 | succeeded | 03d0c373e594aa5715323fd2ba901ecc5a0250f4deb9d9e32371e95dc8a36c27 | 9,417 | null | """abacus: faithful port of the Abacus recipe (McLeish et al., arXiv:2405.17399).
The load-bearing idea is the embedding. Each digit is indexed by its
SIGNIFICANCE WITHIN ITS OWN NUMBER (0 = least-significant), from a table
SHARED across N, x and T. Digit k of N and digit k of x therefore land on the
same embedding an... | {
"id": "046719e6-a33a-4933-9ba3-b44f553b126d",
"created_at": "2026-08-25 16:36:48.166946+00:00",
"db_md5": "80b03c6db927f2e12cf645f2453373df",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "b1bf5934-7d88-47e3-b871-333ffb57cedf",
"tier": "easy",
"dataset_id": "e8",
"status": "succee... | {
"score": {
"mean_loss": 7.313319206237793,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05044563487172127
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.308193683624268,
"example_count": 85,
... |
04696228-6ccf-4236-87b6-42dac25a238a | easy | erdavis0 | 2026-08-27 01:32:44.081631+00:00 | succeeded | dea5a50d1a7e1542b606c505bb2e064069826b4fcc486a84827b8cbefc9b781f | 31,293 | null | """Typed-prefix reversible transition B0: ordinary evaluator token CE.
Public delimiter syntax routes opaque N, X, and T fields into three strict
functional roles. X initializes only an evolving auxiliary tape, N initializes
only an invariant environment lane, and T reaches only a learned monotone
prefix controller. ... | {
"id": "04696228-6ccf-4236-87b6-42dac25a238a",
"created_at": "2026-08-27 01:32:44.081631+00:00",
"db_md5": "9a82ec9520ce6d08ad83d5dbf700b0ab",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "9c242870-2fa8-4467-8fd7-f84148f38800",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 5.546348255028634,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013750000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.650191259524578,
"example_count": 800,... |
046a8c42-9003-4c33-8563-114bc6c1eebf | easy | AFeinfield | 2026-08-19 23:46:10.289075+00:00 | succeeded | 13b18126cac1d39e27cb85c362a5e538b57a3ed7e5fecdb72d0dc4430e26df88 | 19,306 | null | """092: INSTRUMENTED BASE — two measurement fixes, no architecture change.
(1) Step-indexed LR schedule with a fixed screening step budget. The old
schedule decayed on WALL CLOCK, so throughput jitter changed the
LR-per-step trajectory: identical files on e2 gave 1470 vs 1025 steps
and train acc 0.725 vs 0... | {
"id": "046a8c42-9003-4c33-8563-114bc6c1eebf",
"created_at": "2026-08-19 23:46:10.289075+00:00",
"db_md5": "fbfde8069d90ebb1b53792fca27f8bb0",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "9f2c89eb-e31c-4d00-a663-d2c2a8467326",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1512148075157302,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1572561584780567,
"example_count": 60... |
04857e19-da24-4bb6-911b-681bbacdad39 | easy | chad-atexpedient | 2026-08-14 16:24:20.802337+00:00 | succeeded | 70ff666d07233de55f6f19278e3b2210ebda20c1a789b5f95634df8d6fe11e2a | 17,098 | null | """X39: Delta consistency with weight=0.2 (higher than X33's 0.1)."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,... | {
"id": "04857e19-da24-4bb6-911b-681bbacdad39",
"created_at": "2026-08-14 16:24:20.802337+00:00",
"db_md5": "73820e4e54025ce7a7e02969f7e9c67c",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "3e22f5f0-bbec-42af-859e-34623fc64208",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.310756064461848,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3656218586481206,
"example_count": 600... |
04876083-1cc8-40fb-bed1-f8537fe23ab6 | easy | ehonig | 2026-08-19 17:17:38.211866+00:00 | succeeded | fd39a8caa9796ffe701380ad16c9d439ce469ec577bf106ae5168e26a914aa26 | 18,780 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "04876083-1cc8-40fb-bed1-f8537fe23ab6",
"created_at": "2026-08-19 17:17:38.211866+00:00",
"db_md5": "cff6a14fca5bc6e4f02e293d4e1f4221",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "62ded99b-bfb0-4851-a694-c1de96d25402",
"tier": "easy",
"dataset_id": "e2",
"status": "succ... | {
"score": {
"mean_loss": 2.217898368835449,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014375000260770321
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2157328128814697,
"example_count": 300... |
048b2dc3-ea78-4f70-af47-fcc44854d3b7 | easy | sapient-sapiens | 2026-08-15 08:11:28.433786+00:00 | succeeded | cfa2d5ad3daf9c1fbb1288b322672a7bc1a400c5c9d65e165862e656026b9d9e | 11,188 | null | """Narrow/deep looped Transformer with optional latent workspace tokens.
Research candidate for sampled-modulus fixed-T=2. The encoder builds immutable
prompt memory. A stack of learned operator blocks updates a mutable latent
stream, and the whole stack is reused for each task recurrence. Optional blank
workspace ... | {
"id": "048b2dc3-ea78-4f70-af47-fcc44854d3b7",
"created_at": "2026-08-15 08:11:28.433786+00:00",
"db_md5": "73e276376bb5710a1ed15990e994132d",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "a3ae7f90-c037-4e99-af89-0302efaf2630",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.1279412270025784,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.126337704861119,
"example_count": 800,
"e... |
049502f0-8ade-46fc-998a-ea7b323dedf2 | easy | AFeinfield | 2026-08-27 23:22:32.764074+00:00 | succeeded | f21353005ef453c024938830c01022e64188404fc15a36c9baf6e4a1e5c48a56 | 13,899 | null | """161: 159 + PER-STEP STATE CLAMP (+-1e4, generic numerical hygiene).
159/160 failed hosted with NaN: bounded features removed the loss-side
pressure valve on place magnitudes, so iterated squaring inflated the
state to inf and sin(inf)=NaN. Clamping the state after carry
normalization keeps every intermediate finite... | {
"id": "049502f0-8ade-46fc-998a-ea7b323dedf2",
"created_at": "2026-08-27 23:22:32.764074+00:00",
"db_md5": "30d34d0cfa1fa5d77735a1422e252455",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "b69bf15c-0e2a-41f0-8ceb-b7badfca932c",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.157501697540283,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1715309619903564,
"example_count": 100,
... |
049c5794-afb9-417c-8e3d-65cda18e525c | easy | ehonig | 2026-08-20 17:53:45.689088+00:00 | succeeded | 7e90b5e07388cfe377d251dfb149648ef3ad12f1d5c309a5f37dde1a24946403 | 25,168 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "049c5794-afb9-417c-8e3d-65cda18e525c",
"created_at": "2026-08-20 17:53:45.689088+00:00",
"db_md5": "2dcd4232f8c787f5039583b6cb2a3769",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "d5de6395-62e4-4875-9c42-ffaf60b0c08a",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 1.8801151514053345,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8336774110794067,
"example_count": 10... |
049e33c8-10da-469b-ac0c-5c82ce3715d0 | easy | yashkant | 2026-08-14 17:22:06.123261+00:00 | succeeded | 2410d2d297fec9987531c95e0f2985a34d0f33f6f806bf01d9187516486b777c | 34,545 | null | """Round1778 zero-init parsed-T difficulty residual on Round1766 W8/D2.
The proven Round1766 carrier remains exact at initialization. After each
ordinary two-projection Dykstra cycle, a 48-parameter per-coordinate residual
reads the base step, anchor error, correction state, parsed T, cycle phase, and
correction satu... | {
"id": "049e33c8-10da-469b-ac0c-5c82ce3715d0",
"created_at": "2026-08-14 17:22:06.123261+00:00",
"db_md5": "363834dfabf26941977ba7581172a8c0",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "5d332a93-2861-4d33-b1ab-92770045212c",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.8984534740448,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.32125002797693014
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.270968437194824,
"example_count": 300,
... |
04a3a843-5882-4436-a615-7b22b30d5c00 | easy | liam-gb | 2026-08-13 02:21:13.474343+00:00 | succeeded | 5ae4da782d01d42a5e2630f18f3c7dc668f4f397431ff7645051fa3182e49267 | 16,651 | null | """rns_pm_gather_ct_widet_drank: dense loop RETAINED, plus a parallel
residue track supplying the CRT rank.
Base: widet with lookup tables at lr 3e-2.
The wide readout is what makes long answers reachable; the trim is what pays
for it, by not running loop iterations whose update is multiplied by zero.
Base docstring... | {
"id": "04a3a843-5882-4436-a615-7b22b30d5c00",
"created_at": "2026-08-13 02:21:13.474343+00:00",
"db_md5": "1c479f1e07ec8e176359058d7ba96f62",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "a4f3e82a-68b7-4ae8-a550-123337bd83d6",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 9.119710922241211,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.11062500055413693
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.435474395751953,
"example_count": 800,... |
04a4dea3-9a2d-4e7f-8631-2ca42e94c403 | easy | DDanlov | 2026-08-19 20:25:52.520378+00:00 | succeeded | 00b3c70942be83a26d2c0a1efab84fbcf554c1ef89149b893916a3726ff9e13a | 17,727 | null | """
Dynamic Submission for One Layer Deeper API Benchmark
Config: dim=2560, num_heads=32, d_ff_mult=4, fixed_trec=224
"""
from __future__ import annotations
import math
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
... | {
"id": "04a4dea3-9a2d-4e7f-8631-2ca42e94c403",
"created_at": "2026-08-19 20:25:52.520378+00:00",
"db_md5": "8a997de6a91aec208d6cbbd2e7c54d7e",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "161a4702-1d31-4dd0-b522-0e43b920f10e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.9026580605929366,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002083333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9217948181212217,
"example_count": 60... |
04b216b9-fdf9-47b3-af6a-92e4ad68a540 | easy | yashkant | 2026-08-19 01:53:41.506559+00:00 | succeeded | 22b48a4860926826a1d5de2844dbe04a937b1b2a5f2174afc64332e2aa050cf4 | 49,530 | null | """Round2217 adds a zero-start square/reduce/answer recurrence.
The confirmed correction-reset model remains intact. Its width-8
answer state is refined by one shared transition with explicit product,
reduction, and answer roles. Immutable categorical N and X summaries
are reinjected on every call and physical calls... | {
"id": "04b216b9-fdf9-47b3-af6a-92e4ad68a540",
"created_at": "2026-08-19 01:53:41.506559+00:00",
"db_md5": "054a7a506a679f95413f32c3db1665d8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "b4c000ae-4870-492a-96b7-75f4bffebbe3",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.0716502434929844,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016041666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0747673516850833,
"example_count": 12... |
04b993f4-aab3-4cee-a94a-7c9b5f6df36c | easy | chuk-uzowihe | 2026-08-06 23:18:09.396738+00:00 | succeeded | 8fd68575d300ca3e8610d29ca870bc10d94297244c0b8411ac2fbdefb5995133 | 33,265 | null | """One Layer Deeper submission: FPRM over a digit lattice with a Fourier T latent.
Structure (v4)
--------------
1. N and x are *digit slots*: each digit is gathered into a dense per-number
grid (column j = place value 10^j, short rows left-padded with explicit
zero digits) and embedded as digit token + learned ... | {
"id": "04b993f4-aab3-4cee-a94a-7c9b5f6df36c",
"created_at": "2026-08-06 23:18:09.396738+00:00",
"db_md5": "8216ee5199435f3cebf5c8bad57b546c",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "8610aa14-2c3f-4514-b56a-88cb678f88d5",
"tier": "easy",
"dataset_id": "e4",
"status": "... | {
"score": {
"mean_loss": 6.941180250406596,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0034375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.8743688397343385,
"example_count": 1200,
... |
04bdbb86-c6c8-4336-87ec-643cac4d1043 | easy | DDanlov | 2026-08-16 00:16:24.329531+00:00 | succeeded | 8a64ae572e6fd9bd46ea61e2db05ac28203f4362c82cfa5d1946858f7876782d | 16,546 | null | """
Production Port-Hamiltonian RoPE Dynamic Submission
Config ID: 8 | Name: AggressiveEuler_S8_d512_dt020
dim=512, stages=8, heads=16, dt=0.2, lr=0.004, wd=0.1, opt=shampoo
"""
from __future__ import annotations
import math
from typing import Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.function... | {
"id": "04bdbb86-c6c8-4336-87ec-643cac4d1043",
"created_at": "2026-08-16 00:16:24.329531+00:00",
"db_md5": "18561f241db447b37b6dac6da2f3e714",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "dd1c4d5b-fbe6-43c8-b3c1-72da23b0b373",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.171231776696537,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010000000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.181508532554045,
"example_count": 600,... |
04c0a074-b937-4bd3-89d1-bd7b698f0922 | easy | gauravmishra | 2026-08-09 04:28:04.839247+00:00 | succeeded | 683a05c8a1ac6d366537e80f594bd115ba87b23b1c3a863d84971818185a8ed8 | 5,040 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
D_MODEL = ... | {
"id": "04c0a074-b937-4bd3-89d1-bd7b698f0922",
"created_at": "2026-08-09 04:28:04.839247+00:00",
"db_md5": "e3d5156a58ee304d8694783f31aacf87",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "481c7675-aa1f-4d0a-ac33-9d520415bfad",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.8368198129523825,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0037500000000000003
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9568133674929493,
"example_count": 6... |
04c70465-1322-4f97-a32c-a3991d175f6a | easy | AFeinfield | 2026-08-13 01:00:24.300092+00:00 | succeeded | c56bb8d80baf2ee084fa088680c06564d43b52c6b26eb563fac88c709749f59c | 7,736 | null | """Looped (weight-tied) bidirectional Transformer, d=256, 8 iterations.
021: variable-N debut on e3 — 017 (end-pos) with moderated LR (2e-3,
warmup 200) per 019's stability lesson; wd 0.1 (fit-first), grokfast off
to amortize the evaluator's ~200ms/iteration overhead into 8x more steps.
prior. Same parameter count ord... | {
"id": "04c70465-1322-4f97-a32c-a3991d175f6a",
"created_at": "2026-08-13 01:00:24.300092+00:00",
"db_md5": "6da757689e6d5717245bc84921b08b64",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "3d260009-81a6-403b-9f2f-86289d49afcb",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 6.565580188864494,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004375000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.985090513313624,
"example_count": 800,... |
04cb827d-fc18-4df6-8534-702249bd84b6 | easy | oupadhyay | 2026-08-26 15:35:52.240000+00:00 | failed | 8c92d82af4966a77e942149145a0652c072060ad8bfc66a6aabd9e6e178038e2 | 4,046 | null | """Literal polynomial cell."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle,Submission,TokenLossBatch,assert_model_state
D,H=48,24
class C:
def __init__(s,vocab_size,max_seq_len):s.vocab_size,s.max_seq_len=vocab_size,max_seq_... | {
"id": "04cb827d-fc18-4df6-8534-702249bd84b6",
"created_at": "2026-08-26 15:35:52.240000+00:00",
"db_md5": "98c9abd38b118da68d0a7d3c657f11bc",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "868d2ba5-fe1c-4191-998b-de162c1e47ca",
"tier": "easy",
"dataset_id": "e5",
"status": "f... | null
|
04d0f25f-db4c-4bb3-bba0-b14d322a6cc3 | easy | yashkant | 2026-08-16 00:33:53.406604+00:00 | succeeded | 83de680f8c03e86b303fd643c5d3e1086a6328e1de964bb3160a0c55cd37aae2 | 42,377 | null | """Round1801 finite-ring spectral composition inside W12 Dykstra.
The complete Round1765 W12/1T carrier stays active. Each parsed-T cycle maps
the live answer state into a low-capacity real character bank, rotates answer
and immutable X modes by categorical-N phases, and composes answer and latent
frequencies with a ... | {
"id": "04d0f25f-db4c-4bb3-bba0-b14d322a6cc3",
"created_at": "2026-08-16 00:33:53.406604+00:00",
"db_md5": "13c92c04e860bfcfefb4b6f13bf18758",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "bbda6f1b-19e5-4470-95f8-2b23fb9347e9",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1436728677505332,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666679084302
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1476691453446186,
"example_count": 60... |
04d73528-3028-43e7-8658-bd430ff008c8 | easy | gauravmishra | 2026-08-10 08:30:31.397502+00:00 | succeeded | 3634113f9c0009502cc7045676d432ae9172ec6822491d98b339d91a0e39d3e7 | 6,764 | null | """E47: deterministic NoPE on historical K13-S385.
Architecture parent SHA-256:
42dc862219390b8300e98260725c6c2621b27aae32f750f81902dea4bf95889d
NoPE prototype SHA-256:
1033a97da26482afbd31a1e871b7ac4c977988b8a41e28e6a900ae88c95b4f5a
"""
from __future__ import annotations
import torch
import torch.nn.functional as F... | {
"id": "04d73528-3028-43e7-8658-bd430ff008c8",
"created_at": "2026-08-10 08:30:31.397502+00:00",
"db_md5": "044aef8f3ae5f3020ab6cb8d807344ef",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "65a821e8-2f7b-49b1-a62e-eb4f0eefbe65",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.4512233943021324,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5817404178225942,
"example_count": 60... |
04db3b6f-c11f-409e-8d06-bf548526d353 | easy | c-nnr | 2026-08-04 12:49:03.941687+00:00 | succeeded | 88dca25ef8fb8bf4d26bd28ffada22762cf43cbbd1dc10167196a9c323103797 | 5,668 | null | """One Layer Deeper E2 round-two experiment: Smaller faster recurrent cell to trade width for more updates."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
... | {
"id": "04db3b6f-c11f-409e-8d06-bf548526d353",
"created_at": "2026-08-04 12:49:03.941687+00:00",
"db_md5": "17a2d741c3aa39df940e59860196b5af",
"submitter": "c-nnr",
"github_login": "c-nnr",
"run_id": "a22c6906-3287-4394-b340-9c1d21d7e546",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.5624425411224365,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0052083334885537624
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.734835386276245,
"example_count": 30... |
04e1a563-09c0-44c8-baf2-5c720e9ba6fa | easy | yashkant | 2026-08-16 01:07:55.500134+00:00 | succeeded | f4d56a694d397f258dd70a901f14c6fb382285e5711112ce01fc696cfbc40b2f | 32,342 | null | """Round1779 direction-preserving radial Dykstra projection bound.
The proven Round1766 W8/D2 carrier is retained. Every learned-center
projection computes both the parent's coordinatewise tanh map and a radial
soft projection with the same asymptotic W8 vector radius. The candidate
preserves the centered direction ... | {
"id": "04e1a563-09c0-44c8-baf2-5c720e9ba6fa",
"created_at": "2026-08-16 01:07:55.500134+00:00",
"db_md5": "17f00fa563c7b09d383f9e5c5ec3ea87",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "f3d4f479-90ac-44a9-8600-f36249bf876d",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1226264299822764,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011875000158324836
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.121773879368565,
"example_count": 800... |
04ecd0df-be9b-4a1a-b372-f56fa9542ad6 | easy | DDanlov | 2026-08-09 00:03:37.480882+00:00 | failed | c34f65750356624cf02161d9875cf0458ba6d35278f317d346fe9a99853487f6 | 15,981 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "04ecd0df-be9b-4a1a-b372-f56fa9542ad6",
"created_at": "2026-08-09 00:03:37.480882+00:00",
"db_md5": "d7870ffb8d13591e83d47a1dcdc7eb1e",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d71cbb80-d242-41e6-a750-1e05e20f4389",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
04f31763-3e72-4a5a-9525-160b751ee144 | easy | abhishekmittal15 | 2026-08-11 20:39:43.928583+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "04f31763-3e72-4a5a-9525-160b751ee144",
"created_at": "2026-08-11 20:39:43.928583+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Abhishek Mittal",
"github_login": "abhishekmittal15",
"run_id": "063dfb69-f5c1-494c-a28c-8a69da1f5efe",
"tier": "easy",
"dataset_id": "e1",
"sta... | {
"score": {
"mean_loss": 3.0371100902557373,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333334140479565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.97149658203125,
"example_count": 100... |
04f75a21-6cfc-4fea-aedd-828487005360 | easy | erdavis0 | 2026-08-19 09:54:10.101364+00:00 | succeeded | 3a470214c54e81a6fdbb5d2db7b891c47f79af6467c0efdd358d5f25ff77111a | 10,596 | null | """Learned-slot low-rank operator with calibrated endpoint supervision."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLoss... | {
"id": "04f75a21-6cfc-4fea-aedd-828487005360",
"created_at": "2026-08-19 09:54:10.101364+00:00",
"db_md5": "95e07760f3e09e3319abe1e605d1375c",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "b47c0156-a702-405c-8bbe-5abf5854b34e",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.2669733401900123,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016458333830038707
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.265346438488755,
"example_count": 300... |
04f8ff62-d53e-4995-82ca-4213c83f882f | easy | erdavis0 | 2026-08-29 01:08:44.079460+00:00 | succeeded | c30171170d2368108bbd1f8de018dfbd1d387e0e28dec2d543ae91ea35b7ad59 | 20,136 | null | """Curvature table with a GPU-resident cached modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "04f8ff62-d53e-4995-82ca-4213c83f882f",
"created_at": "2026-08-29 01:08:44.079460+00:00",
"db_md5": "efb4da441ead29fd5826581673cba669",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "df2bc72f-fe2f-4d93-8d54-4b4d250dafec",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.54220312833786,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7119369804859161
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.689042806625366,
"example_count": 60,
... |
04fdbc58-764c-4534-bffa-ce412191f4b7 | easy | DDanlov | 2026-08-24 20:45:40.007944+00:00 | succeeded | ab80873f8352b85bc58df0834fc2d407f082cbf11329c1dc14d7eb4726662a88 | 49,569 | null | """
Unified Pure Skipless Prefix Deliberation Model + Transformer Readout Decoder
Architecture:
- Deliberation Core: dim=1536, num_heads=24, d_ff=6144 (Pure Skipless SO(2)^{D/2} Givens Recurrence)
- Token Layout: Recurrent loop processes ONLY [Prompt (L_p), 4 Pause Registers] (Zero target tokens in loop)
- Recurrent At... | {
"id": "04fdbc58-764c-4534-bffa-ce412191f4b7",
"created_at": "2026-08-24 20:45:40.007944+00:00",
"db_md5": "26884e19ba52d2d6a06696577a1d9939",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "089bc78e-d182-4b87-afa3-829d1ecc8b5c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.89377225236258,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000037252902
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8909370204258393,
"example_count": 600,... |
0501d340-ea78-484f-9676-8e55d8b3bb0b | easy | DDanlov | 2026-08-29 13:47:31.580321+00:00 | succeeded | 3e421e48f89ceaf8b3118a7836ead55b93e8ee9122ca712cbc09fb5b4a4381fe | 24,021 | null | """
trial_w07_l3_d768_k12.py
W07: L=3, D=768, H=16, d_ffn=2304, K=12, dt=0.08, mu=0.85, lr=0.0035, freq=5
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
try:
... | {
"id": "0501d340-ea78-484f-9676-8e55d8b3bb0b",
"created_at": "2026-08-29 13:47:31.580321+00:00",
"db_md5": "5a2a0b2782e702eb88dbd6980ed6c4c5",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ba95abb3-6887-4789-88a2-d849010df7a2",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2971498127740024,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012916666703919569
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3329761528113497,
"example_count": 60... |
05032b58-2bb5-4cf7-bf6e-b1de5f07d97b | easy | yashkant | 2026-08-10 21:28:05.398663+00:00 | succeeded | 5009134053ca85753760fce70a90333e1b374205628be5201f1e7dc4459fe898 | 22,265 | null | """Standalone compact-bigram hash, AdamW LR3.75e-3, EMA 0.997."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
from torch import ... | {
"id": "05032b58-2bb5-4cf7-bf6e-b1de5f07d97b",
"created_at": "2026-08-10 21:28:05.398663+00:00",
"db_md5": "d928b459701ca69468b300c96bb056f8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "08d16167-a394-42e3-a15b-65bb0e90a728",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.523406684398651,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3100000284612179
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.526186943054199,
"example_count": 300,
... |
05043b5b-3752-425f-95b8-d350dacd7cd6 | easy | Yalyenea | 2026-08-12 12:52:33.161695+00:00 | succeeded | e740f280e37e9d9b739a24b46281ed93f6d3b15935065d87e028c7ba848d2c51 | 11,548 | null | """Digit-class gated decimal recurrence with wall-clock WSD."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
... | {
"id": "05043b5b-3752-425f-95b8-d350dacd7cd6",
"created_at": "2026-08-12 12:52:33.161695+00:00",
"db_md5": "f9caab073ceff2f6c3935dfb1a937b93",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "97ffbf79-3cf7-4209-8967-7457fb24e638",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.9890660340211657,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01541666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.759799449433126,
"example_count": 600,... |
050645b3-07db-4116-96f2-50596fa4ad9d | easy | DDanlov | 2026-08-18 09:21:02.437244+00:00 | succeeded | 2daa3dabbf47e1f004651c2e7895944312e775b363dd9a9f753435d4e887457f | 23,511 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "050645b3-07db-4116-96f2-50596fa4ad9d",
"created_at": "2026-08-18 09:21:02.437244+00:00",
"db_md5": "a45a27d6e56da3f75f753d1345b5cf43",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "98c28d4c-12bb-4168-9a57-ad11a341e610",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.1827476840516375,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.187113925079594,
"example_count": 600... |
050d0372-f21b-4d95-b4ea-3f6bff60122c | easy | shan23chen | 2026-08-08 03:28:06.825761+00:00 | succeeded | ab9a57b57634c2796d9f979806bc145e46112ba96ccc5c45568857c1d27b8b7d | 25,416 | null | """Boundary-reset recurrent register Transformer.
One learned transition is applied once per requested outer step. The
transition receives an immutable decimal N register and the current learned
boundary state, but never receives T or a separate copy of the original X.
Within each outer application, scratch persists ... | {
"id": "050d0372-f21b-4d95-b4ea-3f6bff60122c",
"created_at": "2026-08-08 03:28:06.825761+00:00",
"db_md5": "b93e656ee236a261c55eb3f5b8266533",
"submitter": "Shan Chen",
"github_login": "shan23chen",
"run_id": "82387456-77a3-44ac-9970-fec655a18504",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.5978686809539795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01000000024214387
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5319085121154785,
"example_count": 600... |
050e9f8b-2f17-4f91-aa0a-9d1307ffb817 | easy | DDanlov | 2026-08-10 08:16:50.188453+00:00 | succeeded | 34d0b3d271411a11f23f01c44a58786a0b44ce76c62edd5041cdc490f5efc88c | 16,397 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "050e9f8b-2f17-4f91-aa0a-9d1307ffb817",
"created_at": "2026-08-10 08:16:50.188453+00:00",
"db_md5": "1d0642f26e76ce078a6d206fddc3825a",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "36ba1e76-26ee-40bf-b45f-281d03680a3f",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 23.18093118405996,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 25.917415618896484,
"example_count": 100... |
050f2e24-ec52-4c2a-a6d8-23726d3dd0d1 | easy | LauraGomezjurado | 2026-08-30 21:40:13.123137+00:00 | succeeded | a547638bd3d7d4d6f32c3db29d84ca7238026ad3cc49ed548523f03bf4ab8895 | 9,044 | null | """Looped depth-recurrent Transformer for repeated modular squaring.
Design notes
------------
The evaluator reads answers off the *last* ``target_len`` prompt positions, so
answers are right-aligned with the least-significant digit at the final valid
position. Digit significance is therefore a fixed offset from the e... | {
"id": "050f2e24-ec52-4c2a-a6d8-23726d3dd0d1",
"created_at": "2026-08-30 21:40:13.123137+00:00",
"db_md5": "436efb6e9a38ae6fbc8a4133f21e320c",
"submitter": "LauraGomezjurado",
"github_login": "LauraGomezjurado",
"run_id": "2b918ddb-ee7c-4402-a5c5-c6e1a03665d4",
"tier": "easy",
"dataset_id": "e7",
"st... | {
"score": {
"mean_loss": 2.4659724874271243,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.032798574084611926
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.485674517329146,
"example_count": 85,... |
0513582f-b2be-4094-a0b4-293d275a63d1 | easy | isaac0804 | 2026-08-07 11:52:57.411571+00:00 | succeeded | 8d6a2a1d3230e4d3854edc4941a5ae1d9cac56af351aab0b0afe432ee28612ae | 10,696 | null | """nGPT: Normalized Transformer with Representation Learning on the
Hypersphere (Loshchilov et al., arXiv:2410.01131), simplified port.
Motivation: a local diagnostic session on E3 (docs/findings.md) found that
the standard RMSNorm pre-LN baseline (`baseline_adamw_deep`) grows an
unbounded residual-stream norm as trai... | {
"id": "0513582f-b2be-4094-a0b4-293d275a63d1",
"created_at": "2026-08-07 11:52:57.411571+00:00",
"db_md5": "24ff3d054d342db94a227516a9cf128a",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "58f606df-9232-4e0b-b4a1-ed37e3c451ff",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 2.1908209648915493,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0050000000093132265
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.190652915094397,
"example_count": 80... |
051906cf-7161-4068-9d18-6617a0f35264 | easy | khushidahi | 2026-08-06 04:41:19.281020+00:00 | succeeded | ed4203ae66c22e137c695e9dfd0a42a23329e4427b58f19cf1d0e7638a2552bc | 14,399 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "051906cf-7161-4068-9d18-6617a0f35264",
"created_at": "2026-08-06 04:41:19.281020+00:00",
"db_md5": "b200d849729cbabf17fc48bfd3c1062c",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e1b4aee8-a8b6-4ea6-81c8-477e8309a320",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.196706175804138,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333507180214
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1954469680786133,
"example_count": 600... |
051d9e1b-c1e8-4c53-9fc1-f88d5ca61cbb | easy | anirudh-chakravarthy | 2026-08-12 03:20:53.712957+00:00 | succeeded | e30a8112ce66fa9ea028723bbb99b3846eb23e6d14a8943fc1eabedc67f59b8b | 14,572 | null | """A learned state transition applied once per requested time step."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
asser... | {
"id": "051d9e1b-c1e8-4c53-9fc1-f88d5ca61cbb",
"created_at": "2026-08-12 03:20:53.712957+00:00",
"db_md5": "71e46dc53df7f6690d0aba9e54723890",
"submitter": "Anirudh S Chakravarthy",
"github_login": "anirudh-chakravarthy",
"run_id": "323091fb-a45d-48fa-aadf-e5aa208825c7",
"tier": "easy",
"dataset_id": "... | {
"score": {
"mean_loss": 3.5499565648978653,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000062088171
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.235662252913676,
"example_count": 600... |
051facaa-aebf-4687-a818-484aa36ba9b5 | easy | gauravmishra | 2026-08-08 20:39:59.691231+00:00 | succeeded | 162030f55db9498d104f2fdb8d3665f8c69ba41136aad58bed20f15436a8e2ec | 4,820 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
D_MODEL = ... | {
"id": "051facaa-aebf-4687-a818-484aa36ba9b5",
"created_at": "2026-08-08 20:39:59.691231+00:00",
"db_md5": "046fae8d8759cb2dedf20ed59274b58b",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "b3257202-7bbd-457a-af57-883f8fb5a93c",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.227746939495858,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2470301876153647,
"example_count": 600... |
05236e1a-780b-4ef8-83ec-7205be979a3d | easy | k-penchev | 2026-08-05 02:55:12.646458+00:00 | succeeded | eb9da847382468a798cce9b2e115719078294f75fbe065479f91c1613844ba50 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "05236e1a-780b-4ef8-83ec-7205be979a3d",
"created_at": "2026-08-05 02:55:12.646458+00:00",
"db_md5": "e95b392799a3b50162b830594a1c7b3f",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "7553944c-a364-4a6f-b579-4b1ab57e299e",
"tier": "easy",
"dataset_id": "e3",
"status": "... | {
"score": {
"mean_loss": 2.426050639772299,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.639098861611687,
"example_count": 800,... |
052c47bd-8289-4758-b9d0-c95353647f97 | easy | adwhit | 2026-08-13 12:46:48.207454+00:00 | succeeded | d0c8edc553969290631cd24f39618eba9b5b6ef7536b772e2de0e9632c55f87d | 15,365 | null | """DeepThinker: depth-recurrent transformer for One Layer Deeper.
Architecture (task-general — no arithmetic is hard-coded):
- Weight-tied recurrent block unrolled to fixed depth (train 24 / eval 70),
so effective serial depth extrapolates beyond training T at evaluation.
- Recall connection: the original toke... | {
"id": "052c47bd-8289-4758-b9d0-c95353647f97",
"created_at": "2026-08-13 12:46:48.207454+00:00",
"db_md5": "d5c48b38c5673036815b7e4228ea94f9",
"submitter": "adwhit",
"github_login": "adwhit",
"run_id": "ff3d273f-5184-48cd-bc38-72fa8f1ebe6f",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.7977755069732666,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.046666666865348816
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.025815963745117,
"example_count": 100... |
052e60ab-77a0-4027-a6a0-d70d505fa3ad | easy | liam-gb | 2026-08-13 02:19:53.420360+00:00 | succeeded | 2e9df07bf71dc171766b494c72f5b91e6381f20e012c1b7a4b542be81aeb6dae | 15,040 | null | """rns_pm_gather_ct_widet_norelrt: tables at 3e-2 AND batch reuse off.
Base: widet with lookup tables at lr 3e-2.
The wide readout is what makes long answers reachable; the trim is what pays
for it, by not running loop iterations whose update is multiplied by zero.
Base docstring:
Single change vs rns_pm_gather_ct:... | {
"id": "052e60ab-77a0-4027-a6a0-d70d505fa3ad",
"created_at": "2026-08-13 02:19:53.420360+00:00",
"db_md5": "a06277be536373d81f9d0a6e692bd1a6",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "3beaebee-e12d-45a1-8607-7148529e8705",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 6.51228336048038,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.050312499863406024
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.857753783360487,
"example_count": 1200,... |
052f2922-3248-4713-95d7-dd1c68cb6c80 | easy | DDanlov | 2026-08-26 01:10:49.509717+00:00 | succeeded | 6ec702e80a3ef8108f77dd779fe8242495c65b2b94998c8676e893e0adcca1f6 | 35,819 | null | import math
import time
from typing import Optional, Tuple, Any, Dict, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
def log_debug(msg):
with open("debug_log.txt", "a") as f:
f.write(msg + "\n")
try:
from benchmark.api import ModelSpe... | {
"id": "052f2922-3248-4713-95d7-dd1c68cb6c80",
"created_at": "2026-08-26 01:10:49.509717+00:00",
"db_md5": "2dcdab405c8297c0c60df6fea24ac6f2",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c542f222-b4c5-4f70-a305-4027caef37a1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3062412119576825,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.304544166599154,
"example_count": 600,
... |
05321ab9-4960-41df-a13a-6cfce5e74ac3 | easy | chad-atexpedient | 2026-08-13 01:54:22.643731+00:00 | succeeded | d68219b8969eb07ccbd4a6ae14fb9bc795a33557e6c527f97f8e3094b4dce15b | 13,856 | null | """X17B: pair behavior with CE."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
LANES = 6
C... | {
"id": "05321ab9-4960-41df-a13a-6cfce5e74ac3",
"created_at": "2026-08-13 01:54:22.643731+00:00",
"db_md5": "f65ee37e691a991067f07e66d497b4a3",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "88ff444e-d160-46df-a73d-f0d7894d5589",
"tier": "easy",
"dataset_id": "e3",
"st... | {
"score": {
"mean_loss": 2.167901130871348,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00937500006519258
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.16949341966166,
"example_count": 800,
... |
0535193c-6eb2-4ae2-8a74-4708223156dd | easy | shirvani-jr | 2026-08-29 02:48:43.373381+00:00 | succeeded | 459da493dd2908cad9a8c9872fbad420462944f11d66ff0b4106740ae7ffd3e9 | 47,739 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "0535193c-6eb2-4ae2-8a74-4708223156dd",
"created_at": "2026-08-29 02:48:43.373381+00:00",
"db_md5": "bd4bfcc15bf65f683c4a1d57d23e89cb",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "28a9238e-6674-41a0-a537-dfb17fde4c47",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 0.633395828306675,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8616666495800018
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.1609676629304886,
"example_count": 100,
... |
0538cb7e-f26e-4c16-b367-48d7bc9c500c | easy | sapient-sapiens | 2026-08-16 15:40:22.713981+00:00 | succeeded | 9e3ddca97c21fb4055b571c1dab84c547b4291348a92bdde23885f1965068f4b | 23,310 | null | """E6 generalization candidate: width-384 ACT64 shared recurrence.
Architecture changes relative to the hosted ACT max-4 source:
- selectable recurrence depths are 1 through 64;
- canonical cumulative-mass halting genuinely removes halted examples from
later recurrent applications;
- the T-field representation pres... | {
"id": "0538cb7e-f26e-4c16-b367-48d7bc9c500c",
"created_at": "2026-08-16 15:40:22.713981+00:00",
"db_md5": "6a1c6d30b24564c6b5fa98d367a3d321",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "dce86ced-8e31-4634-bb50-8e9067d17e13",
"tier": "easy",
"dataset_id": "e6",
"status": "su... | {
"score": {
"mean_loss": 5.931737739745884,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.09189189487212412
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.254183769226074,
"example_count": 60,
... |
053edeff-ca72-4748-b740-f36aca2cfa8e | easy | Alan-Qin | 2026-08-18 17:17:19.093051+00:00 | succeeded | 0bc5d50fb167839ac0b3414e658dfef674352a761e2d3917c11b42c088a3bbec | 16,654 | null | """One-hot CRT loop + sibling-consistency supervision (v24).
Architecture: per-band one-hot residues (parameter-free encoding of x),
a learned per-band transition matrix W_m applied T times (STE one-hot), and a
shared digit decoder read at every loop step.
Training loss (token_training_loss): besides the final-answer... | {
"id": "053edeff-ca72-4748-b740-f36aca2cfa8e",
"created_at": "2026-08-18 17:17:19.093051+00:00",
"db_md5": "cf3600bf05b023d27d45e55095750f32",
"submitter": "Zeyu Qin",
"github_login": "Alan-Qin",
"run_id": "43097ae1-ee3c-4618-8a40-8548faff6a5c",
"tier": "easy",
"dataset_id": "e8",
"status": "succeede... | {
"score": {
"mean_loss": 0.23342212289571762,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9461675882339478
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.20579062402248383,
"example_count": 85... |
053ee0ce-7bc3-4c2e-9185-4bf6788a5f49 | easy | rougee | 2026-08-22 06:45:23.672443+00:00 | failed | 782e3558cf806f580cfd0fe774885fe3fe3d368703d8f33a41fae5fecf23bb9f | 980 | null | import torch
import torch.nn as nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission
class MinimalModel(nn.Module):
def __init__(self, spec: ModelSpec):
super().__init__()
self.config = spec
self.embed = nn.Embedding(spec.vocab_size, 64)
self.layer = nn.TransformerEncoderLay... | {
"id": "053ee0ce-7bc3-4c2e-9185-4bf6788a5f49",
"created_at": "2026-08-22 06:45:23.672443+00:00",
"db_md5": "2b647565a9b8526d23b9062c442e5359",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "8447f7f1-f4e7-41c4-a8e4-77a35131d4ed",
"tier": "easy",
"dataset_id": "e6",
"status": "failed",
"... | null
|
0540c632-6bd2-4b32-9614-2d2173ce33b7 | easy | yunjiangster | 2026-08-26 17:15:45.421956+00:00 | failed | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "0540c632-6bd2-4b32-9614-2d2173ce33b7",
"created_at": "2026-08-26 17:15:45.421956+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "1648d471-b7f4-434c-8ff6-66a1dab70593",
"tier": "easy",
"dataset_id": "e6",
"status":... | null
|
0552a0a7-f690-4ab3-8386-457771041abb | easy | DDanlov | 2026-08-15 16:08:20.866461+00:00 | failed | e464973e16b0ba4603430f8e107737ccdfb945d93b54f48524beaeea88280290 | 18,629 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Model B - Scaled Continuous Lie Group Commutator Core (LieCommutatorNet)
- Skew-Symmetric Lie Algebra Generators in so(1024) + Commutator Tensor Network
- dim=1024, num_generators=32, d_hidden=2048, RohanShampoo with lambda=0.12
"""
from __future__ i... | {
"id": "0552a0a7-f690-4ab3-8386-457771041abb",
"created_at": "2026-08-15 16:08:20.866461+00:00",
"db_md5": "4288ed9d246c22280bf20b4d2ce82a9f",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "9c373cb3-7d03-4401-88bc-3563f5d94976",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
05559805-7a37-4c4d-a5e9-6bae1bfc12b6 | easy | TZZheng | 2026-08-06 21:15:20.336923+00:00 | succeeded | a7f51e1b4f35491d7245225398862605760974d0531f191864e1aa06cbe98239 | 7,301 | null | """Self-contained standard Transformer scaling-sweep submission."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
VARIANT = {'block... | {
"id": "05559805-7a37-4c4d-a5e9-6bae1bfc12b6",
"created_at": "2026-08-06 21:15:20.336923+00:00",
"db_md5": "0640470de367f227b4d66e618db11928",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "9ee2f097-bade-4bc4-acea-6b593201c52c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.476080060005188,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06999999657273293
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.0309321880340576,
"example_count": 100,... |
055d2ba5-04f5-45e7-b260-68472ec30124 | easy | erdavis0 | 2026-08-17 12:14:49.101452+00:00 | failed | ea190f18bd9a305e9f0613da12c85ea70f8688cff6c04a1d5a872582d1cc459e | 20,800 | null | """Minimal sign-tied transition with integer-turn local regression."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatc... | {
"id": "055d2ba5-04f5-45e7-b260-68472ec30124",
"created_at": "2026-08-17 12:14:49.101452+00:00",
"db_md5": "14823b4a802bd51a370881679403f199",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "54a1d7ef-d69b-45bf-9963-4e1b9a51b7d3",
"tier": "easy",
"dataset_id": "e7",
"status": "failed",
... | null
|
055eb0db-fdd2-4e2a-b686-aa4575ce13e1 | easy | karanganesan | 2026-08-07 07:05:33.322644+00:00 | succeeded | 28ca8eb014e5334d6f2b7616acf9c6e75aec12217230fa80579e0a6947c43625 | 25,610 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "055eb0db-fdd2-4e2a-b686-aa4575ce13e1",
"created_at": "2026-08-07 07:05:33.322644+00:00",
"db_md5": "633edcb54cef5aa0bcd7b0ba92deb99b",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "b2a1f241-5790-495a-8e13-e99a9f7c5fba",
"tier": "easy",
"dataset_id": "e4",
"status": ... | {
"score": {
"mean_loss": 2.921486973762512,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008229166734963655
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.608964443206787,
"example_count": 1200... |
055f106b-3d0d-485c-a36a-721f9f701e57 | easy | yashkant | 2026-08-15 11:54:41.510259+00:00 | succeeded | 56c58ff04a92d5bd9a48a440254e01c0fea296a76379d2c18c5d52fc96c92184 | 45,947 | null | """Round2001 dual-timescale EMA merge on frozen Round1939.
The live Round1939 architecture, initialization, sparse-SAM8 optimizer, and
learning-rate schedule are unchanged. Both matched arms store and update the
same .995 and .999 EMA banks after committed optimizer updates. At evaluation
they eagerly form the fast-on... | {
"id": "055f106b-3d0d-485c-a36a-721f9f701e57",
"created_at": "2026-08-15 11:54:41.510259+00:00",
"db_md5": "470bdef721f901c5badd76e25f4b1c2b",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "ba939b97-9734-4cf9-9506-0424c5579979",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.068810833900293,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0104166666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0713543220129425,
"example_count": 1200,... |
0566b203-15de-4618-af6e-abba15e3f861 | easy | yashkant | 2026-08-15 14:03:14.112468+00:00 | succeeded | 03b304d52edd6ab0187de662ff3eecd25efca4637bc588d916c6022f98bb2af0 | 33,286 | null | """Round1786 evolutionary crossover of two controlled survivors.
The Round1780 W8/D2 carrier keeps its winning contrastive, exactly
antisymmetric learned set codes. The candidate additionally removes each
initial and recurrent Dykstra correction residual's coordinate mean before
the unchanged tanh bound, combining Ro... | {
"id": "0566b203-15de-4618-af6e-abba15e3f861",
"created_at": "2026-08-15 14:03:14.112468+00:00",
"db_md5": "267de64e1cd8a82c539c970185491b70",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "6c7853ee-a88e-4de3-82bf-69e88ca67d9e",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.901716411113739,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2762500233948231
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.051044464111328,
"example_count": 300,
... |
0570a108-7c18-43b9-ab52-c3ed9b326a77 | easy | jordanrubin | 2026-08-05 19:20:43.967079+00:00 | succeeded | 952a66e2c24b667b7a190e8c6465a29280facd54861d8ad96628b6ccd1efa4e0 | 9,327 | null | """L1 continuous loop with a parameter-free RMS boundary projection.
The one-block cell is reused exactly T times. Before its output becomes the
next logical state, every token is projected to unit RMS. This removes
residual-stream norm drift without adding a second block, an optimizer state,
or an output bottleneck... | {
"id": "0570a108-7c18-43b9-ab52-c3ed9b326a77",
"created_at": "2026-08-05 19:20:43.967079+00:00",
"db_md5": "1b14e83969bdd54b49b3192223f07617",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "739d84de-2275-43fd-bb32-ef1649eb2cf1",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 4.808982034719571,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.09999999947225054
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.314751879817241,
"example_count": 300,
... |
0576b5d1-0876-4e2d-983d-51b6faa3a194 | easy | sankalp1999 | 2026-08-04 12:06:29.402401+00:00 | succeeded | 16346c71573718bcb6e285f0cc5eedbaf9d2ed7ce0c179562b5b6fd450471a4e | 7,815 | null | """Latent recurrence: encode once, apply a shared squaring map T times, decode.
v4 put the loop inside a transformer and underfit badly (train acc 4% after
4,432 steps) because attention cost ~7ms/step at 13 tokens and bought nothing.
v2 fit the train set perfectly with a cheap flattened encoder but memorized,
because... | {
"id": "0576b5d1-0876-4e2d-983d-51b6faa3a194",
"created_at": "2026-08-04 12:06:29.402401+00:00",
"db_md5": "404dbc8139944d2d5490c95adcc5bfa0",
"submitter": "Sankalp ",
"github_login": "sankalp1999",
"run_id": "7b80bf6c-8259-403c-a3f0-5a7f45c85cd7",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 8.102104663848877,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06666666828095913
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.123236656188965,
"example_count": 100,... |
057e5e1e-443c-4ab9-9ead-eb25015bb3c4 | easy | viridale | 2026-08-04 21:10:18.984483+00:00 | succeeded | 56d92c95c9fc7fcb74a78ecb7f4e61b236850b1d9684d93fa30b8f72f39765f8 | 8,194 | null | """Clean token-native tied-recurrent baseline after the Rule-10 reset.
hypothesis: a generic sequence encoder plus a fixed-depth shared neural refinement
block is the strongest defensible zero point after removing the manually supplied
field parser, prime/factor axes, residue/CRT state, exact recurrence catalo... | {
"id": "057e5e1e-443c-4ab9-9ead-eb25015bb3c4",
"created_at": "2026-08-04 21:10:18.984483+00:00",
"db_md5": "19013d4fa0c2eb669ddb9be2b063d172",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "7a99de80-1fc6-400b-82e4-c610b0422bec",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 2.582222231143337,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009375000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5937461415709295,
"example_count": 800... |
0582a97d-7e02-4532-a3ec-7745058cfa2d | easy | richardcepka | 2026-08-13 20:05:35.510576+00:00 | succeeded | 2917f3b12226ca18d4323550c170ecb27d524d6509129864c8f0aac9e8e861d3 | 20,134 | null | """Recurrent digit-register model for One Layer Deeper.
A looped transformer whose recurrent state is a decimal digit register that is
re-quantised through a ten-entry codebook on every iteration. Nothing inside
the loop depends on the iteration index and no parameter is indexed by an
absolute slot position, so the e... | {
"id": "0582a97d-7e02-4532-a3ec-7745058cfa2d",
"created_at": "2026-08-13 20:05:35.510576+00:00",
"db_md5": "a0ee77fb704b9d894ce30448e0e14939",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "89b3c008-752e-4498-b448-6dbb68e478ad",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.14629340714617,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1480538139428793,
"example_count": 600,... |
0584b1a7-223c-4959-967b-ad5446bf1fdc | easy | shirvani-jr | 2026-08-28 00:54:33.982341+00:00 | succeeded | 396fee419635addaf93a12d3a1446ceda5cdae5e6fca6dbe436e0fa97b7929b8 | 48,549 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "0584b1a7-223c-4959-967b-ad5446bf1fdc",
"created_at": "2026-08-28 00:54:33.982341+00:00",
"db_md5": "6eecdcb811fc7c9142f7f4a70d9f9fa1",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "2b5478b4-9ef3-41fc-97bb-6607c9220eb1",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3840922117233276,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05133689986541867
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.538891077041626,
"example_count": 85,
... |
058a0d9e-056a-4ab4-8c0e-a7f30e425b23 | easy | karanganesan | 2026-08-08 05:24:58.905220+00:00 | succeeded | 8d9eecdc8607c91b190336e24b49597738decc69ce7d4e24ee5930ae97cfcad7 | 27,127 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "058a0d9e-056a-4ab4-8c0e-a7f30e425b23",
"created_at": "2026-08-08 05:24:58.905220+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "8e18d10b-43f4-477d-8d5a-efbe308a558d",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 7.095664739608765,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666795499623
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.467165470123291,
"example_count": 600,... |
058c2233-fb74-4951-9836-11e4c04c1701 | easy | yashkant | 2026-08-09 02:47:23.498510+00:00 | succeeded | 9084788c57d70e9371c16efd7b28d9e381cd80dfe176a2de867c67d09ded40c6 | 22,836 | null | """Round 211 common TRM-style x/y/z shared-transition template."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
from torch import... | {
"id": "058c2233-fb74-4951-9836-11e4c04c1701",
"created_at": "2026-08-09 02:47:23.498510+00:00",
"db_md5": "3e558755dcaae2ce6d98f73727e41002",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "76d6c288-2dff-4e4e-8a88-34dea117536b",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 4.310199975967407,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05833333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.753617286682129,
"example_count": 100,
... |
0591c365-abbf-4b24-a57a-4637e00c3081 | easy | sapient-sapiens | 2026-08-19 19:02:50.506314+00:00 | succeeded | faeba505df0d7dfc9aa507497f814b275ef227cb5ac6415906d3a733bb351c56 | 32,507 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "0591c365-abbf-4b24-a57a-4637e00c3081",
"created_at": "2026-08-19 19:02:50.506314+00:00",
"db_md5": "6ab21e535799c06f298ad8aad73855be",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "6210059b-3a15-49da-94f9-2d33e855e91a",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.535514206562876,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5540365296094407,
"example_count": 600... |
059c1e24-86c0-472c-8df2-b698b3204ec3 | easy | jordanrubin | 2026-08-08 09:58:53.998544+00:00 | succeeded | 078cd8887c504bdcfcf580d442de465beb5b3a0ea5e1726a6c95fd2267731c91 | 16,615 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "059c1e24-86c0-472c-8df2-b698b3204ec3",
"created_at": "2026-08-08 09:58:53.998544+00:00",
"db_md5": "1f66d0c9d8b2addfecdef9a9aa694dae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "87f5ca41-4529-4aa2-8550-71c221fc9450",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1699986189893665,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01374999969266355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1771992942142915,
"example_count": 600... |
059edcb2-c996-4c7d-92a0-06861a67bbb7 | easy | sapient-sapiens | 2026-08-31 22:00:51.167558+00:00 | succeeded | 7d3da506e1c5e52e41aa5c4b5b9e4d28324e447964ac8be968358b41492b6974 | 16,043 | null | """One-block nested E5 model: four learned Horner calls per square, eight squares."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_mod... | {
"id": "059edcb2-c996-4c7d-92a0-06861a67bbb7",
"created_at": "2026-08-31 22:00:51.167558+00:00",
"db_md5": "4f0b86636fcfea84df34200cb2b622e5",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "0f06d0ea-c4d9-44f3-bb7d-bf6a9d6e3b08",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1401146952343826,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1367845246610084,
"example_count": 60... |
05a30bc7-1a3d-445a-8b8a-19b0d6b05d8a | easy | KaustubhKumar05 | 2026-08-26 11:22:08.297687+00:00 | succeeded | 8db2c3f5053f5be5e9d1a22c90e1c7bdb39310e69c46e962abf6ea96a689e235 | 11,393 | null | """sq_gru — candidate A2: digit-canonical recurrence with a measured-good step.
Built from components that MEASURED well, not from theory:
* E9b encoder (FINDINGS 30): embed digits -> bidirectional 2-layer GRU -> decode
every output digit. On held-out operands drawn from the same range it squares
small numb... | {
"id": "05a30bc7-1a3d-445a-8b8a-19b0d6b05d8a",
"created_at": "2026-08-26 11:22:08.297687+00:00",
"db_md5": "273d3ba49e0d55af91969c16220116fe",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "9b27ad31-318b-4a0f-be0d-db54c1fa77cf",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 2.28555365178513,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2913640146832828,
"example_count": 600,... |
05a51ff4-ddbf-4b42-a1a2-6b6c799e210c | easy | khushidahi | 2026-08-12 20:30:39.584050+00:00 | succeeded | 7ab0e1a944bf4d67c7829d4d7d60f68645e4d31d4ecd649d08f80f78abc7ba95 | 24,154 | null | """R53 learned digit-pair table with directional carry/reduction scans.
The model learns a full 10x10 pair embedding table from final-label training.
For the diagonal variants, pair feature (i,j) is routed to decimal place i+j.
No multiplication values, carry rules, modular-reduction rules, generated
examples, or inte... | {
"id": "05a51ff4-ddbf-4b42-a1a2-6b6c799e210c",
"created_at": "2026-08-12 20:30:39.584050+00:00",
"db_md5": "6f66e250032b9ea21d854c71501912e8",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "28373815-3efd-4750-b9a0-83b69bc77342",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.3031901121139526,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666800156236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3216822147369385,
"example_count": 60... |
05a74eb7-2ced-494a-82e0-cf1bc0b09985 | easy | erdavis0 | 2026-08-22 12:02:37.442440+00:00 | succeeded | 51c1ac6f1c1b63d62814de1d34409c49ecd688b4e0533ed3562b45bbe4363796 | 9,114 | null | """Iterative latent workspace for soft prompt-role discovery and slot readout."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_s... | {
"id": "05a74eb7-2ced-494a-82e0-cf1bc0b09985",
"created_at": "2026-08-22 12:02:37.442440+00:00",
"db_md5": "8c1df270c6f26cdb1236360125a41b1a",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "2369583c-478f-4a49-91db-130f1bbe6637",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 6.66647481918335,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.10757575929164886
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.3801116943359375,
"example_count": 125,
... |
05a77897-9755-4e0e-a265-5c4b0e7b0a8a | easy | sapient-sapiens | 2026-08-20 15:24:58.006764+00:00 | succeeded | 197de4aeda0f330c0ee8f00dfc4a6e97e226d061fc9d0aaff35555cb8ab0e289 | 35,002 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "05a77897-9755-4e0e-a265-5c4b0e7b0a8a",
"created_at": "2026-08-20 15:24:58.006764+00:00",
"db_md5": "04348a9093a67cc9627440f886babf6e",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "091bfcf1-497d-43a7-b376-47d7429ca626",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.7140380975953278,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.792004383198349,
"example_count": 600,
"... |
05a88449-676f-49ac-ae3c-862c14ca4975 | easy | blake-camp-surge | 2026-08-04 02:49:02.942158+00:00 | succeeded | 8cedb24bd9a4a88dffdcbe335edf0b088fd801759ce45ca528a7bea1a1b0f508 | 10,023 | null | """Vector-state virtual neurons driven by one shared QKV-generating DAN."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
... | {
"id": "05a88449-676f-49ac-ae3c-862c14ca4975",
"created_at": "2026-08-04 02:49:02.942158+00:00",
"db_md5": "1d958bdd816891b8f96bfc589925fdd3",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "e45fc21c-ad7e-4637-8155-887134319264",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 2.1425560116767883,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4043490886688232,
"example_count": 100... |
05a97848-f1a2-427d-a797-7c5ec9e9e1e4 | easy | RobertTLange | 2026-08-03 09:39:32.224823+00:00 | succeeded | e5cf9b39e8e1f12d4f839c3730e633d6cfe6dbdd2f374f0aef85939e03b0a9a3 | 36,410 | null | """A weight-tied recurrence whose *depth* is a latent variable, not a guess.
The target is repeated modular squaring, so the trajectory a solver should walk
is ``x, x^2, x^4, x^8, ...`` -- one squaring per step, and completely
independent of the requested ``T``. Only the *stopping point* depends on ``T``.
Every ances... | {
"id": "05a97848-f1a2-427d-a797-7c5ec9e9e1e4",
"created_at": "2026-08-03 09:39:32.224823+00:00",
"db_md5": "c495cba20e86dbf78c343241b52767ba",
"submitter": "Robert Tjarko Lange",
"github_login": "RobertTLange",
"run_id": "2ec82b89-fa14-4cc0-a251-d4911ef21bef",
"tier": "easy",
"dataset_id": "e1",
"sta... | {
"score": {
"mean_loss": 1.8465183973312378,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06833333149552345
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.7952017784118652,
"example_count": 100... |
05aa831b-73c3-4433-9d92-af32125a5765 | easy | yashkant | 2026-08-14 11:04:38.085207+00:00 | succeeded | 882409b73583024c5f534f37a403a2301c0a145ddd2b2bafb06ac29bf63f6e2a | 20,745 | null | """Round1956 Round493 block-two shared-head supervision.
The exact Round493 model, initialization, optimizer, and final prediction path
remain intact. During training both arms additionally read out the hidden state
after encoder block two with the already-existing norm, attention pool,
decoder, and answer head. The c... | {
"id": "05aa831b-73c3-4433-9d92-af32125a5765",
"created_at": "2026-08-14 11:04:38.085207+00:00",
"db_md5": "44bd6c90bb314307b60ffde74f35f1ac",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "d3daa735-322d-44e9-9661-428f2081d452",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 5.1581199169158936,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08999999985098839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.474495887756348,
"example_count": 100,... |
05bc669e-2abc-4ac7-92ec-c82e4857e5e2 | easy | velocizapkar | 2026-08-30 16:37:19.629410+00:00 | succeeded | 5c8ee47bbc801493bd7ba5eb52143381406519f108fd1d2986fa986067fa1ac7 | 7,486 | null | """HeightGrid: a (sequence x depth x height) transformer for One Layer Deeper.
The computation is a 3D grid. Each cell (t, d, h) holds a width-W vector and
receives three interactions:
* sequence: softmax attention over cells (., d, h) in the same plane;
* height: the running state ascending from (t, d, h-1);
... | {
"id": "05bc669e-2abc-4ac7-92ec-c82e4857e5e2",
"created_at": "2026-08-30 16:37:19.629410+00:00",
"db_md5": "60f12ad7eb595dacf72033c30dfd11dd",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "f0a969d6-fa5c-4890-89f5-5b4cdef98a7c",
"tier": "easy",
"dataset_id": "e1",
"statu... | {
"score": {
"mean_loss": 2.0902093052864075,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04333333298563957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.267692804336548,
"example_count": 100,... |
05c6603d-d6a3-4a9f-88ed-4e6a8534dda1 | easy | gauravmishra | 2026-08-10 03:44:44.900585+00:00 | succeeded | 83fd13930151f22c5c317e4ee15aa6e9906aaacfb966322fc611c6043905a746 | 32,464 | null | """Information-gain candidate ig60_e28_moe4_e5 (independent)."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
C... | {
"id": "05c6603d-d6a3-4a9f-88ed-4e6a8534dda1",
"created_at": "2026-08-10 03:44:44.900585+00:00",
"db_md5": "daa5bf88c46fac224b1d4c7a428a6988",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "5b0c60fb-7776-46be-b35c-bc35796f240a",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.413241773742329,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4686604294541707,
"example_count": 600,
"... |
05ca9865-0015-45cf-bdcb-09bc152313ab | easy | DDanlov | 2026-08-23 21:47:10.421626+00:00 | succeeded | bfbccc3d81aacecea71f0530adec894681ba5e95007bd1af400d3b2f168f6efa | 31,834 | null | """
Unified Hybrid Masked Deliberation Model (71.9M Parameters) - Tier 1 (Pure Skipless, No Cross-Attention)
Architecture:
- Scale: dim=2048, num_heads=32, d_ff=8192 (Exactly 71,889,920 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target Toke... | {
"id": "05ca9865-0015-45cf-bdcb-09bc152313ab",
"created_at": "2026-08-23 21:47:10.421626+00:00",
"db_md5": "0627aec424101548b27b053a1e674e77",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "b3509a0c-d400-4491-8c4d-692364252f24",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.4614794727038447,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.4413563969958525,
"example_count": 600,
"exa... |
05ccb001-e55b-408e-bba1-afbd74f625b7 | easy | chad-atexpedient | 2026-08-09 02:45:13.121894+00:00 | succeeded | ba3e2165647b484e79d9f9a0869c2a33f8292f9e6eb508ce1df1f68290be98e6 | 17,960 | null | """X04B: categorical pair/feedback factorial arm.
The mutable register is initialized from x once. Every subsequent application
of the tied cell receives only the current LSD-first digit state and immutable
N digits. Requested T controls only the number of applications. The state
logits are also the answer logits,... | {
"id": "05ccb001-e55b-408e-bba1-afbd74f625b7",
"created_at": "2026-08-09 02:45:13.121894+00:00",
"db_md5": "fe177da68ad08f396523e9dd8c26d8d6",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "dfacfaf3-938d-4f15-8e21-bd614242ae89",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.8600367938948645,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333221574625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5520007610321045,
"example_count": 60... |
05d6e891-3fcc-48df-8b62-6a28ef6bbf75 | easy | KaustubhKumar05 | 2026-08-26 02:57:27.387509+00:00 | succeeded | 6bc9f8a286cef3811c312c4440a52a842e6772ebc2d944142432682045bc2f1e | 12,204 | null | """Neural-GPU-style conv GRU, v2s: v2 + speedrun pack.
- LR: warmup, hold max to 85% of wall-clock, steep late cliff (2026 schedule theory)
- identity-at-init: cell output projection zero-initialized
- batch reuse x6 (multi-pass theory)
Base (v2):
Changes over v1 (each backed by measured results in the literature):
... | {
"id": "05d6e891-3fcc-48df-8b62-6a28ef6bbf75",
"created_at": "2026-08-26 02:57:27.387509+00:00",
"db_md5": "2e50eea62c768830a8538730226b0e0d",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "8b05b527-03c6-4e55-896b-157d6d669ef1",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 1.5009996891021729,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7054054141044617
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.39752840995788574,
"example_count": 60,... |
05d750b4-430d-4945-b0f8-c2a3ae70ea0c | easy | liam-gb | 2026-08-16 20:37:34.430681+00:00 | succeeded | 24fc878b19987564e88eccf1d98937ae5735fa8ab620fdef470b9eacfec14e2f | 16,740 | null | """rns_pm_gather_ct_widet_l32nore_r8: single change vs
rns_pm_gather_ct_widet_l32nore — the batch-reuse policy returns True, so the
evaluator reuses each fetched batch for the maximum eight optimizer updates.
WHY. On the granular Easy suites one batch is 85-97% of the training split,
so every step ends an epoch and th... | {
"id": "05d750b4-430d-4945-b0f8-c2a3ae70ea0c",
"created_at": "2026-08-16 20:37:34.430681+00:00",
"db_md5": "46c5decfff7d97a850fb297981e9d310",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "94f7a79c-b852-45ed-8815-f32e5fc7b2ba",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.9078171849250793,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2709447480738163
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.5546793937683105,
"example_count": 85,
... |
05e54359-e765-4bce-977b-e1a5f4c4d2cf | easy | shirvani-jr | 2026-08-29 01:18:38.791784+00:00 | succeeded | 4ff1a358e6b860871d2258e5da18e2073b7971cb606f43655a33247a493192a5 | 48,764 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "05e54359-e765-4bce-977b-e1a5f4c4d2cf",
"created_at": "2026-08-29 01:18:38.791784+00:00",
"db_md5": "c21828f9572ed69144a851e9621a9378",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "eaf1df18-3e55-4688-a8ec-04cfa327d26c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 0.8349658846855164,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8300000131130219
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.3931475877761841,
"example_count": 100,... |
05e6513c-f2a9-4335-9ade-a5af541fd8c8 | easy | DDanlov | 2026-08-18 10:52:19.223176+00:00 | succeeded | 9e2e886a7f72218c6f6788250c79c4edcba3908cfb09a02021dbdad2941008ba | 23,584 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "05e6513c-f2a9-4335-9ade-a5af541fd8c8",
"created_at": "2026-08-18 10:52:19.223176+00:00",
"db_md5": "d6174d2a9211082098f18de894eca987",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "b69bfeec-e24d-4682-a2ca-e65fea069e41",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.270133876849524,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.26873954615274,
"example_count": 800,
"e... |
05eb6dda-efd0-498c-8f78-7ec2a7ddf73c | easy | DDanlov | 2026-08-30 00:22:39.385655+00:00 | failed | a3e3bf4613448d60efa17d3e1605361dc3b141a70b485276ddd7da4630f31053 | 28,233 | null | """
Unified Continuous Hopfield Equilibrium Propagation Deliberation Engine
=======================================================================
Architecture:
- Deliberation: Continuous Hopfield Multi-Layer Energy System
- Integration: In-Place Strang-Split Conformal Symplectic Velocity Verlet (with LaSalle Resets)
... | {
"id": "05eb6dda-efd0-498c-8f78-7ec2a7ddf73c",
"created_at": "2026-08-30 00:22:39.385655+00:00",
"db_md5": "64d8df42302b646af7d65aa53bea99c0",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c81af96e-1d5d-40d2-be88-1eb7946e896b",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
05eb95f4-e495-48b5-b7df-e3da7177b16a | easy | blake-camp-surge | 2026-08-05 19:04:37.487452+00:00 | succeeded | 3b81b82d8449da77d1699a0fb0feae33d75fa4567ec242ef6e285f64bb849bf7 | 10,518 | null | """Vector-state neural cellular automaton over attention edges."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
TOKEN... | {
"id": "05eb95f4-e495-48b5-b7df-e3da7177b16a",
"created_at": "2026-08-05 19:04:37.487452+00:00",
"db_md5": "5eb4db4302f4964d7b718376e7701d25",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "e104d892-2079-47c6-8881-da625b0e6d07",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 2.162097806786244,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1580127788081507,
"example_count": 800... |
05f53878-397a-49eb-bdd9-64d294bbfd7b | easy | Khanference | 2026-08-15 20:03:48.621245+00:00 | failed | d3c7e3735203fe416b900ccd5dd28e017929b8be2010d97ad804edead7d4a238 | 4,916 | null | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerSpec,
OptimizerBundle,
Submission,
TokenLossBatch,
assert_model_state,
BatchReuseContext,
)
D_MODEL = 1024
RHO = 4
MAX_STEPS = 100000
BATCH_SIZE = 512
EVAL_BATCH_SIZE... | {
"id": "05f53878-397a-49eb-bdd9-64d294bbfd7b",
"created_at": "2026-08-15 20:03:48.621245+00:00",
"db_md5": "8beaf1a871554ef4cc7c8b1f10e9a84f",
"submitter": "Khanference",
"github_login": "Khanference",
"run_id": "d1f4a891-4dfb-4cc6-9629-481450170de0",
"tier": "easy",
"dataset_id": "e1",
"status": "fa... | null
|
05f91f2b-61e5-4f91-b6fc-6407d5ec0125 | easy | sdrshn-nmbr | 2026-08-05 23:40:20.211446+00:00 | succeeded | 00e1b0ec51095afca533ddede35c0a94d55b88407ac492156b9a7cbf2de6f491 | 54,740 | null | """Typed model and optimizer factory for controlled architecture experiments."""
from dataclasses import dataclass, replace
from functools import partial
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
BatchReuseContext,
ModelSpec,
... | {
"id": "05f91f2b-61e5-4f91-b6fc-6407d5ec0125",
"created_at": "2026-08-05 23:40:20.211446+00:00",
"db_md5": "ed757e9f14e52f66e824a47afecf8bb8",
"submitter": "Sudarshan Nambiar",
"github_login": "sdrshn-nmbr",
"run_id": "fbbc22a1-d56d-4101-a2cc-bd2444a64492",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 3.4160568714141846,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08499999716877937
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.497049331665039,
"example_count": 100,... |
05f9f783-45fa-4485-abd4-a4273fe995cd | easy | DDanlov | 2026-08-29 13:11:35.381948+00:00 | succeeded | 91ed5f25ede1465db33f0517715aeb03225c9ee51c96d276de38b08bdf433af6 | 24,081 | null | """
trial_05_precond_k24.py
T05: Extreme Preconditioned HeavyBall (K=24, dt=0.06, mu=0.88) on Wide L=2 (D=384).
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
... | {
"id": "05f9f783-45fa-4485-abd4-a4273fe995cd",
"created_at": "2026-08-29 13:11:35.381948+00:00",
"db_md5": "00fa36844313451ddaccf439734a49c6",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "a3aee404-9dbd-4c82-a8d2-fd14af67d176",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.294711255961424,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666679084302
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.328369014867218,
"example_count": 600,... |
05fe2e5b-c76f-437f-8fec-7068a47cfc29 | easy | MarcoF1 | 2026-08-13 20:14:03.645051+00:00 | succeeded | 902b04244e1887e72cce9c5d9e4ac0d2ff0fd77404243b8aac7a1715801f6725 | 17,095 | null | """Population-of-clocks: K independent residue-clock models trained in
parallel as batched weight tensors, ensembled at evaluation.
Rationale: hosted training is kernel-launch-bound, not FLOP-bound, so running
K=8 models costs almost no wall-clock. Each member gets its own init and its
own table-dropout probability (s... | {
"id": "05fe2e5b-c76f-437f-8fec-7068a47cfc29",
"created_at": "2026-08-13 20:14:03.645051+00:00",
"db_md5": "8cc1ab9a4e343d83df4dfd4b204378e6",
"submitter": "Marco Fleming",
"github_login": "MarcoF1",
"run_id": "01e9a70b-f9cc-4172-87f6-f0db8e2556c6",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 3.7862799506532285,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.14333333226541678
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.577481746673584,
"example_count": 100,... |
05ffe726-7b8c-4e18-acb3-6b4cb1f28135 | easy | jordanrubin | 2026-08-08 10:33:55.583951+00:00 | succeeded | 19e434c7e9fc9d9d601292a0f5c0a4083813bc7790caf87c0582fc1bbcc2cad8 | 16,909 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "05ffe726-7b8c-4e18-acb3-6b4cb1f28135",
"created_at": "2026-08-08 10:33:55.583951+00:00",
"db_md5": "671c15c2db39c821999bae96a7defa51",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "6e53a5fc-6303-4b09-83dc-d227083a66dc",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.624768089174662,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007499999832361937
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6542005208842956,
"example_count": 600... |
060196e6-25ec-46a9-8638-4aa403471160 | easy | Dandandan | 2026-08-07 03:17:38.708957+00:00 | succeeded | 27aa3f363eced4a2035efa852a82e475f80c15d0f6718f73b11e8d6fc0803bb6 | 99,457 | null | """Generalized exact recursive learner for One Layer Deeper.
The learned transition is a population of ordinary modular neural units with
several generic arithmetic activations. A single learned evidence selector
chooses one transition, which is then reused for every input-requested outer
step. The population covers ... | {
"id": "060196e6-25ec-46a9-8638-4aa403471160",
"created_at": "2026-08-07 03:17:38.708957+00:00",
"db_md5": "239adce2f605a19fe2ff28b049bcb9f3",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "1bc7a72c-3cdf-49a7-9722-1ce11558f69d",
"tier": "easy",
"dataset_id": "e1",
"status": "suc... | {
"score": {
"mean_loss": 145.24848175048828,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.046666666865348816
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 138.71641540527344,
"example_count": 10... |
0604eb86-0302-4aea-b9ce-1eb438c53a3f | easy | anirudh-chakravarthy | 2026-08-17 21:57:58.330153+00:00 | succeeded | 6f4b8f4d495fd951b5503e3c074f3d85e8e27696e0d74debdf22c4dabc01d0a5 | 18,404 | null | """Wave 1: a T-blind atomic CGRU with uniform or cyclic phase adapters."""
from __future__ import annotations
import math
from typing import NamedTuple
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submiss... | {
"id": "0604eb86-0302-4aea-b9ce-1eb438c53a3f",
"created_at": "2026-08-17 21:57:58.330153+00:00",
"db_md5": "b471ef4746218056009e1d060a42ae2e",
"submitter": "Anirudh S Chakravarthy",
"github_login": "anirudh-chakravarthy",
"run_id": "77493392-9ca0-4924-bf2f-2f969eaa8506",
"tier": "easy",
"dataset_id": "... | {
"score": {
"mean_loss": 1.8659687638282776,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.045270273461937904
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8426352739334106,
"example_count": 60... |
06094861-c891-409b-9b1b-bfd78ebaca3d | easy | oupadhyay | 2026-08-26 07:58:15.239431+00:00 | succeeded | 78adf09a1144990896e2249cb3b31b7f85a8794e9550666f384bd15a5d08a4d7 | 4,091 | null | """Persistent dynamic-width cell with a literal learned polynomial update."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle,Submission,TokenLossBatch,assert_model_state
D,H=32,16
class C:
def __init__(s,vocab_size,max_seq_len)... | {
"id": "06094861-c891-409b-9b1b-bfd78ebaca3d",
"created_at": "2026-08-26 07:58:15.239431+00:00",
"db_md5": "872db13f6397d0fe2de0e5c31c40a43a",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "e840b76d-3f74-448a-b028-97c71d287f12",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.4511369314293487,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.471705442052251,
"example_count": 600... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.