From Noisy Retrieval to Decision Evidence: Silver Evidence Demonstrations for In-Context Learning

A method called silver-evidence demonstrations improved over equally sized gold-only in-context learning on three of four datasets and raised evidence F1 on all annotated datasets, according to an arXiv paper (arXiv:2610.32128v1). The approach replaces raw retrieved passages in ICL examples with LLM-generated, target-conditioned evidence summaries produced without access to the expert label or gold evidence, then pairs that silver evidence with the expert label; at test time a single structured call generates silver evidence from retrieved passages before predicting the label and decision evidence. The method also beat InstructRAG-style ICL on mean label accuracy across EI 2.0, SciFact, synthetic policy retrieval, and HICRIC/IMR data, suggesting compact silver evidence helps decisions over long, noisy text.

arXiv:2610.32128v1 Announce Type: new Abstract: RAG and retrieval-augmented few-shot ICL can struggle when decision-relevant evidence is buried in long, noisy context. Retrieved passages and demonstrations may contain redundant or distracting information. Although expert-labeled development examples may identify the text supporting each decision, these annotations are typically used only as answer-side rationales rather than as supervision for filtering noisy retrieval. We propose silver-evidence demonstrations, which replace raw retrieved passages in ICL examples with generated, target-conditioned evidence summaries. For each labeled development example, an LLM generates silver evidence without access to the expert label or gold evidence; this evidence is then paired with the expert label and gold evidence where available. At test time, a single structured call generates silver evidence from retrieved passages and then predicts the final label and decision evidence. Across EI 2.0, SciFact, synthetic policy retrieval, and HICRIC/IMR data, our full method improves over equally sized gold-only ICL label-only for policy on three of four datasets and improves evidence F1 on all annotated datasets. It also achieves higher mean label accuracy than InstructRAG-style ICL on all four datasets. These results suggest that replacing demonstration-side retrieved passages with compact, target-conditioned silver evidence is particularly useful for decisions based on long and noisy text, where it reduces distracting context while preserving decision-relevant signals and expert supervision.