Iterative self-rewarding via LLM-as-Judge in DPO training on Llama 2 70B improves instruction following and self-evaluation, outperforming GPT-4 on AlpacaEval 2.0.
Rui Yang, Ruomeng Ding, Yong Lin, Huan Zhang, and Tong Zhang
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
MedAgentBench-v3 shows capability ceilings and format-knowledge barriers limit pure RL to 18.2% while rule-based SFT reaches 34.1% on clinical protocol tasks.
REINFORCE self-training on competitive programming tasks exhibits robust rise-then-collapse in pass@1; CARE, ES, and GRPO mitigate it in model-size-dependent ways across Qwen-2.5-3B/7B and a Gemma pilot.
UNA unifies binary, pairwise, and score-based feedback for LLM alignment via a generalized implicit reward function shown optimal by the log sum inequality.
FEST improves RLVR sample efficiency on math and coding benchmarks by combining supervised signals, on-policy signals, and decaying weights on just 128 randomly chosen demonstrations, matching full-dataset baselines.
Derives SimPO from MaxEnt RL and reports that MaxEnt RL in online RLHF exhibits frequent overoptimization and unstable KL dynamics across scales, unlike stable KL-constrained baselines.
PoliLegalLM, trained with continued pretraining, progressive SFT, and preference RL on a legal corpus, outperforms similar-scale models on LawBench, LexEval, and a real-world PoliLegal dataset while staying competitive with much larger models.
citing papers explorer
-
World Feedback for Clinical Agents: Diagnosing RL in FHIR Environments
MedAgentBench-v3 shows capability ceilings and format-knowledge barriers limit pure RL to 18.2% while rule-based SFT reaches 34.1% on clinical protocol tasks.
-
Self-Improvement Can Self-Regress: The Rise-and-Collapse Failure Mode of LLM Self-Training
REINFORCE self-training on competitive programming tasks exhibits robust rise-then-collapse in pass@1; CARE, ES, and GRPO mitigate it in model-size-dependent ways across Qwen-2.5-3B/7B and a Gemma pilot.
-
UNA: A Unified Supervised Framework for Efficient LLM Alignment Across Feedback Types
UNA unifies binary, pairwise, and score-based feedback for LLM alignment via a generalized implicit reward function shown optimal by the log sum inequality.
-
Boosting Reinforcement Learning with Verifiable Rewards via Randomly Selected Few-Shot Guidance
FEST improves RLVR sample efficiency on math and coding benchmarks by combining supervised signals, on-policy signals, and decaying weights on just 128 randomly chosen demonstrations, matching full-dataset baselines.
-
Failure Modes of Maximum Entropy RLHF
Derives SimPO from MaxEnt RL and reports that MaxEnt RL in online RLHF exhibits frequent overoptimization and unstable KL dynamics across scales, unlike stable KL-constrained baselines.
-
PoliLegalLM: A Technical Report on a Large Language Model for Political and Legal Affairs
PoliLegalLM, trained with continued pretraining, progressive SFT, and preference RL on a legal corpus, outperforms similar-scale models on LawBench, LexEval, and a real-world PoliLegal dataset while staying competitive with much larger models.