Leandro von Werra PRO

lvwerra's activity

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

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

topic: win-rate runnable position-swap check

2

#299 opened 1 day ago by

hf-dwarez

topic: capability benchmarks runnable pass@k check

2

#300 opened about 21 hours ago by

hf-dwarez

New activity in rl-llm-wiki/knowledge-base about 7 hours ago

source: arxiv:2103.03874 — MATH: Measuring Mathematical Problem Solving

5

#168 opened 3 days ago by

lvwerra

source: arxiv:2302.04761 — Toolformer: LMs Can Teach Themselves to Use Tools

4

#183 opened 3 days ago by

lvwerra

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

topic: length-bias runnable length-control check

2

#301 opened about 10 hours ago by

hf-dwarez

topic: iterate rl-for-math-and-code — bring current with the 2025 RLVR wave (recipes, data frontiers, elicit-vs-expand)

2

#277 opened 1 day ago by

lvwerra

topic: iterate preference-reward-models — multi-attribute regression reward models (SteerLM + HelpSteer/HelpSteer2)

2

#252 opened 2 days ago by

lvwerra

topic: rl-training-stability-in-practice — weave in PPO-max (Secrets-I) + entropy mechanism

6

#292 opened 1 day ago by

hf-dwarez

topic: rollout-generation-infra — colocated resharding engine + generator layout (clean reopen of #271)

5

#290 opened 1 day ago by

hf-dwarez

source: arxiv:2405.01481 — NeMo-Aligner (clean reopen of #272)

5

#291 opened 1 day ago by

hf-dwarez

source: arxiv:2209.00626 — The Alignment Problem from a Deep Learning Perspective

4

#214 opened 2 days ago by

lvwerra

source: arxiv:2501.01821 - SDPO

4

#296 opened 1 day ago by

cmpatino

New activity in rl-llm-wiki/knowledge-base about 22 hours ago

fix: rlaif — RLAIF (2309.00267) + Self-Rewarding (2401.10020) are now in corpus (de-stale OQ/§6/§7)

2

#295 opened 1 day ago by

lvwerra

New activity in rl-llm-wiki/knowledge-base about 24 hours ago

topic: iterate reasoning-emergence — fold ProRL into §5 (the boundary-expansion counter-position)

2

#294 opened 1 day ago by

lvwerra

fix: dpo-variants — restore ΨPO notation + em-dashes (Unicode regressed by #297)

2

#298 opened 1 day ago by

lvwerra

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

topic: algorithms/dpo-variants - add SDPO

2

#297 opened 1 day ago by

cmpatino

topic: NEW algorithms/self-improvement-and-self-play — method-family hub (STaR/SPIN/Self-Rewarding/Absolute-Zero/TTRL)

2

#286 opened 1 day ago by

lvwerra

topic: iterate reasoning-emergence — fold in the 2025 created-vs-surfaced cluster (pass@k boundary, spurious rewards, self-play)

2

#246 opened 2 days ago by

lvwerra

source: arxiv:2405.01470 — WildChat: 1M ChatGPT Interaction Logs in the Wild

2

#256 opened 2 days ago by

lvwerra

topic: iterate test-time-and-rl-interplay — test-time compute as the training signal (TTRL)

2

#275 opened 1 day ago by

lvwerra