DRSR: Learning Set-Level Deletion Risk for Efficient Long-Horizon Agents
arXiv:2609.27276v1 Announce Type: new Abstract: Long-horizon language-model agents accumulate reasoning traces, tool exchanges, and observations whose relevance changes with the current decision. Existing compression strategies often score historical units independently, but the safety of deleting several units is generally not determined by their singleton scores: redundant evidence, accumulated…
Read original article on cs.AI updates on arXiv.org →