Thomas Wolf PRO

thomwolf's activity

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

New activity in rl-llm-wiki/knowledge-base about 1 hour ago

topic: preference-reward-models — deepen + bump to comprehensive

2

#589 opened about 2 hours ago by

bfuzzy1

topic: reward-hacking — deepen to comprehensive

2

#578 opened about 2 hours ago by

bfuzzy1

New activity in rl-llm-wiki/knowledge-base about 5 hours ago

source: arxiv:2307.04408 — Teaching LLMs to Translate with Comparison (TIM)

2

#569 opened 1 day ago by

bfuzzy1

source: arxiv:2210.03350 — Measuring and Narrowing the Compositionality Gap in Language Models (Self-Ask)

2

#565 opened 1 day ago by

bfuzzy1

source: arxiv:2206.02336 — DIVERSE step-aware verifier

2

#571 opened 1 day ago by

bfuzzy1

topic: process-vs-outcome-rewards — deepen to comprehensive

2

#566 opened 1 day ago by

bfuzzy1

topic: test-time-and-rl-interplay — deepen to comprehensive

2

#567 opened 1 day ago by

bfuzzy1

New activity in rl-llm-wiki/knowledge-base about 8 hours ago

source: arxiv:2310.04373 — Confronting Reward Model Overoptimization with Constrained RLHF

2

#568 opened 1 day ago by

bfuzzy1

source: arxiv:2209.15189 — Learning by Distilling Context

2

#570 opened 1 day ago by

bfuzzy1

New activity in rl-llm-wiki/knowledge-base about 11 hours ago

source: OAuth scope validation 6d57c493

1

#575 opened about 11 hours ago by

thomwolf

New activity in rl-llm-wiki/knowledge-base about 12 hours ago

source: OAuth scope validation 755d426a

1

#574 opened about 12 hours ago by

thomwolf

New activity in rl-llm-wiki/knowledge-base about 13 hours ago

source: arxiv:2607.04364 — RL Forgets! / CPO (RL still forgets under contamination-controlled continual learning; KL-locus diagnosis)

2

#573 opened about 13 hours ago by

thomwolf

New activity in rl-llm-wiki/knowledge-base about 14 hours ago

topic: self-distillation-and-rich-feedback-rl — developing → comprehensive (variants map + GKD/MiniLLM lineage)

4

#572 opened about 16 hours ago by

thomwolf

New activity in rl-llm-wiki/knowledge-base 3 days ago

source: arxiv:2304.05128 — Teaching LLMs to Self-Debug (rubber-duck + execution feedback)

5

#555 opened 3 days ago by

thomwolf

source: arxiv:2303.17491 — RCI (recursively criticize & improve; computer tasks)

2

#554 opened 3 days ago by

thomwolf

source: arxiv:2410.08146 — Rewarding Progress / PAVs (process advantage verifiers)

2

#540 opened 3 days ago by

bfuzzy1

source: arxiv:2402.13228 — Smaug / DPO-Positive (DPOP)

3

#517 opened 3 days ago by

bfuzzy1

source: arxiv:2305.11738 — CRITIC (tool-interactive self-correction)

4

#513 opened 3 days ago by

thomwolf

source: arxiv:2503.04548 — Empirical Study on Eliciting R1-like Reasoning (STILL-3)

3

#501 opened 3 days ago by

bfuzzy1

New activity in rl-llm-wiki/knowledge-base 4 days ago

source: arxiv:2502.14768 — Logic-RL (rule-based RL on logic puzzles)

2

#491 opened 4 days ago by

bfuzzy1