内容库 / 论文

arXiv:2608.24764 · Agent 智能体

Evidence Blindness in Direct Corpus Interaction: Persistent Navigation with AtlasNav

类型：arxiv
标识：2608.24764
链接：http://arxiv.org/abs/2608.24764v1
主分类：agent
形态：method
TLDR：Large language model agents are moving beyond conventional retrieval-augmented generation toward direct interaction with external corpora. Direct Corpus Interaction (DCI) keeps the full corpus accessible, yet reachable evidence can remain unusable under finite interaction budgets. Required evidence may fail to surface, a surfaced supporting document may remain unopened, or an opened document may fail to expose its decisive fragment. We call this progressive silent loss Evidence Blindness and quantify it through stage-wise evidence realization. Within the DCI paradigm, raw interaction adds litt
待LLM分类：否
来源文件：
/inbox/tom/_candidates/2026-08-26-agent-rag-longcontext-candidates.json

相关 · 按语义匹配

仓库 prakash-kharvi/metaverse-screening-record

仓库 Chiagoziem2/Systematic-review-screening

笔记 Tom 文献雷达 · Agent + RAG + Long Context · 2026-08-01

仓库 jlucasmcrell/ComfyUI-H3-Multishot

笔记 Tom 文献雷达 · Agent + RAG + Long Context · 2026-08-01 14:40 CST

笔记 Tom 文献雷达 · Agent RAG LongContext · 2026-08-12 20:40

arXiv

2608.24764

TLDR

Large language model agents are moving beyond conventional retrieval-augmented generation toward direct interaction with external corpora. Direct Corpus Interaction (DCI) keeps the full corpus accessible, yet reachable evidence can remain unusable under finite interaction budgets. Required evidence may fail to surface, a surfaced supporting document may remain unopened, or an opened document may fail to expose its decisive fragment. We call this progressive silent loss Evidence Blindness and quantify it through stage-wise evidence realization. Within the DCI paradigm, raw interaction adds litt

arXiv ↗ 中文全文