Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results
Retrieval-augmented large language models frequently face contexts that interleave useful evidence with misleading statements or instruction-like content. Blanket refusal discards valid evidence, whereas uncritical adoption yields incorrect or unsafe answers. The ability to selectively adopt relevant information while rejecting deceptive or harmful content is therefore critical for reliable deployment in real-world retrieval settings. We introduce SelectBench, a controlled benchmark and training set for selective evidence adoption, and post-train Qwen3.5-4B directly with DAPO using either dete
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- FuzzySimilar title/name (fuzzy) · 59%aymericdamien/TopDeepLearning →
“Fuzzy title match (0.73): “Reinforcement Learning for Large Language Model Selective Ev” ≈ “aymericdamien/TopDeepLearning””
- LinkedLinked via arxiv author · 85%Yanyu Chen →
“Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results”
- LinkedLinked via arxiv author · 85%Tai-Yue Li →
“Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results”
- LinkedLinked via arxiv author · 85%Yongyi Cui →
“Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results”
- LinkedLinked via arxiv author · 85%Dongsheng Shi →
“Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results”
- LinkedLinked via arxiv author · 85%Lichang Dai →
“Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results”