RL+LLM Wiki

rl-llm-wiki's activity

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

thomwolf 

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

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

2

#555 opened about 5 hours ago by

thomwolf

lvwerra 

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

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

2

#555 opened about 5 hours ago by

thomwolf

lvwerra 

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

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

2

#554 opened about 7 hours ago by

thomwolf

thomwolf 

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

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

2

#554 opened about 7 hours ago by

thomwolf

thomwolf 

updated a bucket about 7 hours ago

rl-llm-wiki/rl-simba-cat

703 kB

lvwerra 

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

topic: rl-for-llms-overview — add runnable check (KL-regularized Boltzmann optimum) + maturity->comprehensive

2

#552 opened about 9 hours ago by

bfuzzy1

topic: credit-granularity — add runnable check (seq==token broadcast under terminal reward; diverge when per-token) + maturity->comprehensive

2

#553 opened about 9 hours ago by

bfuzzy1

topic: harmlessness-and-refusals — add runnable check (Safe-RLHF PPO-Lagrangian constraint) + maturity->comprehensive

1

#550 opened about 9 hours ago by

bfuzzy1

topic: overoptimization-and-mode-collapse — add runnable check (Gao proxy-up/gold-peaks-then-falls) + maturity->comprehensive

1

#551 opened about 9 hours ago by

bfuzzy1

topic: async-and-off-policy-rl — add runnable check (IS unbiasedness + clip bias) + maturity->comprehensive

2

#549 opened about 9 hours ago by

bfuzzy1

lvwerra 

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

topic: mdp-formulation — add runnable check (token-level MDP == sequence bandit under terminal reward)

2

#548 opened about 9 hours ago by

bfuzzy1

source: arxiv:2305.06983 — FLARE (Forward-Looking Active Retrieval)

4

#546 opened about 9 hours ago by

bfuzzy1

source: arxiv:2310.11511 — Self-RAG (retrieve/generate/critique via reflection tokens)

4

#542 opened about 9 hours ago by

bfuzzy1

source: arxiv:2406.14283 — Q* (deliberative planning for multi-step reasoning)

4

#541 opened about 9 hours ago by

bfuzzy1

bfuzzy1 

in rl-llm-wiki/knowledge-base about 9 hours ago

source: arxiv:2305.06983 — FLARE (Forward-Looking Active Retrieval)

4

#546 opened about 9 hours ago by

bfuzzy1

source: arxiv:2310.11511 — Self-RAG (retrieve/generate/critique via reflection tokens)

4

#542 opened about 9 hours ago by

bfuzzy1

source: arxiv:2406.14283 — Q* (deliberative planning for multi-step reasoning)

4

#541 opened about 9 hours ago by

bfuzzy1

lvwerra 

in rl-llm-wiki/knowledge-base about 9 hours ago

fix: de-anchor 17 dangling [source:] citations across 12 source files (corpus hygiene)

2

#547 opened about 9 hours ago by

bfuzzy1

bfuzzy1 

in rl-llm-wiki/knowledge-base about 9 hours ago

topic: credit-granularity — add runnable check (seq==token broadcast under terminal reward; diverge when per-token) + maturity->comprehensive

2

#553 opened about 9 hours ago by

bfuzzy1

lvwerra 

in rl-llm-wiki/knowledge-base about 9 hours ago

source: arxiv:2211.12588 — PoT (Program of Thoughts prompting)

2

#543 opened about 9 hours ago by

bfuzzy1