The First Drop of Ink: Nonlinear Impact of Misleading Information in Long-Context Reasoning

ArXi:2605.10828v1 Announce Type: new As large language models are increasingly deployed in retrieval-augmented generation and agentic systems that accumulate extensive context, understanding how distracting information affects long-context performance becomes critical. Prior work shows that semantically relevant yet misleading documents degrade performance, but the quantitative relationship between the proportion of distractors and performance remains unstudied.