Search papers, labs, and topics across Lattice.
This paper investigates whether an agent's recent behavior can predict when context compaction will negatively impact performance in long-horizon tasks. By analyzing the TRACE paired-replay corpus, the study replays compaction boundaries under both pre-compaction and summarized contexts to measure subsequent action burdens (errors or repetitions). The findings indicate that pre-boundary history offers only a weak predictive signal for post-compaction harm, with the best interpretable trigger identifying 21% of harmful boundaries while preserving 84% of compaction opportunities.
Context compaction in long-horizon agents is surprisingly insensitive to recent agent behavior, with simple "has-written" heuristics measuring trajectory phase rather than predictive power.
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.
