Ben

benfielding's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a paper about 2 months ago

The Role of Feedback Alignment in Self-Distillation

Paper • 2606.11173 • Published Jun 9 • 3

upvoted a collection about 2 months ago

Open 1B

Collection

A 1.6B decoder-only LM whose training can be independently verified, with midtrained and SFT variants. Record: https://open1b.gensyn.ai • 4 items • Updated Sep 23 • 5

upvoted 2 papers 5 months ago

F-TIS: Harnessing Diverse Models in Collaborative GRPO

Paper • 2605.22537 • Published May 21 • 13

DEI: Diversity in Evolutionary Inference for Quality-Diversity Search

Paper • 2605.27130 • Published May 26 • 16

upvoted 2 papers 7 months ago

Backdoor Attacks on Decentralised Post-Training

Paper • 2604.02372 • Published Mar 31 • 11

Training-Free Dynamic Upcycling of Expert Language Models

Paper • 2603.29765 • Published Mar 31 • 10

upvoted a paper 12 months ago

Hail to the Thief: Exploring Attacks and Defenses in Decentralised GRPO

Paper • 2511.09780 • Published Nov 12, 2025 • 29

upvoted a paper about 1 year ago

Sharing is Caring: Efficient LM Post-Training with Collective RL Experience Sharing

Paper • 2509.08721 • Published Sep 10, 2025 • 665

upvoted 2 collections over 1 year ago

RL Swarm

Collection

RL Swarm is an open source system for peer-to-peer gossip-based reinforcement learning over the internet. • 5 items • Updated Apr 30, 2025 • 7

Gensyn Papers

Collection

12 items • Updated Jun 10 • 9

upvoted 5 papers over 1 year ago

All is Not Lost: LLM Recovery without Checkpoints

Paper • 2506.15461 • Published Jun 18, 2025 • 41

SkipPipe: Partial and Reordered Pipelining Framework for Training LLMs in Heterogeneous Networks

Paper • 2502.19913 • Published Feb 27, 2025 • 6

Verde: Verification via Refereed Delegation for Machine Learning Programs

Paper • 2502.19405 • Published Feb 26, 2025 • 9

HDEE: Heterogeneous Domain Expert Ensemble

Paper • 2502.19385 • Published Feb 26, 2025 • 5

NoLoCo: No-all-reduce Low Communication Training Method for Large Models

Paper • 2506.10911 • Published Jun 12, 2025 • 9