🔄 In a Training Loop

Asankhaya Sharma

codelion

hugging-science

·

http://asankhaya.github.io/

AI & ML interests

Creator of OptiLLM, OpenEvolve, Adaptive Classifier, and Ellora. Pioneering a new category in AI infrastructure: inference-time compute for LLMs.

Recent Activity

updated a dataset about 8 hours ago

mlx-community/mlx-model-explorer-data

updated a dataset about 8 hours ago

LocalLLaMA/local-model-explorer-data

new activity 2 days ago

LocalLLaMA/typed-decisions:ProKope-421M (Alan Canto): 0.714 acc / 0.071 Brier / 0.254 Score MAE on official test split

View all activity

Organizations

upvoted a collection 3 days ago

DataInventor datasets

Synthetic datasets built with DataInventor: describe the data you want and get training rows back. • 9 items • Updated 3 days ago • 1

upvoted a paper 4 days ago

EvoDuet: Bilevel Co-Evolution of Web Searching and Task Solving for Scientific Discovery

Paper • 2609.40340 • Published 5 days ago • 107

upvoted a collection 6 days ago

OpenFineTrain

Very fine 55 trillion+ tokens worth of ungated huggingface datasets for LLM training across 80+ datasets. • 90 items • Updated 9 days ago • 9

upvoted 2 articles 6 days ago

Article

AI for Food Allergies

hugging-science

•

Oct 16, 2025

• 33

Article

TorchSim: A new PyTorch-based molecular dynamics engine

Nekkrad

•

Oct 31, 2025

• 14

upvoted a collection 7 days ago

Typed decision models (System One)

Open models that answer yes/no, choice and score questions with probabilities, plus typed-decisions, a public benchmark to compare them. • 17 items • Updated 6 days ago • 3

upvoted an article 20 days ago

Article

Training mRNA Language Models Across 25 Species for $165

OpenMed

•

Mar 31

• 30

upvoted 2 papers 24 days ago

LoongFlow: Directed Evolutionary Search via a Cognitive Plan-Execute-Summarize Paradigm

Paper • 2512.24077 • Published Dec 30, 2025 • 5

Automated Discovery Has No Universally Superior Harness

Paper • 2607.18235 • Published Jul 20 • 1

upvoted a paper about 1 month ago

H3-World: Turning Language Understanding into World Control

Paper • 2609.01560 • Published Sep 1 • 53

upvoted 2 papers about 2 months ago

Improving the matrix multiplication exponent with modern optimization and AlphaEvolve

Paper • 2608.16884 • Published Aug 17 • 20

WorldClaw: Agentic 3D Open-World Generation at Scale

Paper • 2608.05248 • Published Aug 5 • 86

upvoted an article about 2 months ago

Article

The Distillation Research: Why 15,000 Clean Traces Outperform 100,000 Raw Ones

Siddh07ETH

•

Aug 9

• 2

upvoted 3 papers 3 months ago

Automated Verification of Integer Overflow

Paper • 1909.09324 • Published Sep 20, 2019 • 1

The Dynamics of Software Composition Analysis

Paper • 1909.00973 • Published Sep 30, 2019 • 1

Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks

Paper • 2606.29082 • Published Jun 27 • 44

upvoted an article 3 months ago

Article

SPROG-9M: how far a 9-million-parameter, LLM-free model gets on grade-school math

codelion

•

Jun 26

• 3

upvoted a paper 3 months ago

Qwen-AgentWorld: Language World Models for General Agents

Paper • 2606.24597 • Published Jun 23 • 167

upvoted 2 papers 4 months ago

Even with AI, Bijection Discovery is Still Hard: The Opportunities and Challenges of OpenEvolve for Novel Bijection Construction

Paper • 2511.20987 • Published Nov 26, 2025 • 1

GigaEvo: An Open Source Optimization Framework Powered By LLMs And Evolution Algorithms

Paper • 2511.17592 • Published Nov 17, 2025 • 122