Self-Compacting Language Model Agents
AI Classified by Officially
Long agent traces composed of chains of thought and tool calls accumulate stale content that anchor subsequent generations, and eventually outgrow the context window. Existing scaffolds mitigate it with fixed-interval compaction triggered at a token threshold. Such triggers pay no heed to trajectory structure, risking discard of partial results mid-derivation or mid-search. …
This is an extract. The publication continues at the source.
Read the original at the source: https://machinelearning.apple.com/research/self-compacting-language-agents
Officially imported this from Apple Machine Learning Research’s own source and shows an extract. If you work there, claiming the profile and verifying the domain lets you choose to show the full text here.
Provenance
- Organization
- Apple Machine Learning Research — imported from official source
- Official source
- https://machinelearning.apple.com/rss.xml RSS
- Imported
- October 13, 2026 00:01
- Versions
- 1 recorded
- Identity
-
self-compacting-language-agents