Agent 历史能否预测上下文压缩的伤害？TRACE 配对重放语料上的有限效应

View PDF HTML (experimental)

Abstract:Many long-horizon agents compact their context on a global rule, usually a token budget, blind to what the agent was doing. We ask whether the agent's recent behaviour predicts when a compaction will hurt. TRACE's public corpus of 590 harness-triggered AppWorld compaction boundaries replays each boundary from a re-executed prefix state under the pre-compaction context and under the summary, and records the burden of the next actions: calls that error or repeat a call already made. We find that pre-boundary history predicts post-compaction harm only weakly. An internally prespecified contrast by prefix placement is a wide null, and the naive "has-written" label behind it turns out to measure trajectory phase. The best extension-protocol trigger reaches held-out AUROC 0.66 (0.64 on the replicate's own label) against a same-boundary replicate of 0.72; the best frozen, interpretable trigger avoids 21% of harmful (positive-burden) boundaries while keeping 84% of compaction opportunities, and exceeds the random-rule expectation on count but not on burden mass (a post hoc comparison). Whether the best trigger beats a token-budget rule at matched retention cannot be evaluated on the release. We state what corpora should ship to answer it.

Comments:	Accepted at the IAB Workshop (Interpreting Agent Behavior) at NeurIPS 2026 (non-archival). 20 pages
Subjects:	Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as:	arXiv:2610.08722 [cs.AI]
	(or arXiv:2610.08722v1 [cs.AI] for this version)
	https://doi.org/10.48550/arXiv.2610.08722 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Egor Pakhomov [view email]
[v1] Tue, 6 Oct 2026 17:26:57 UTC (68 KB)