Metraipolis › archive › dc_1987eca5f458

TAP: Efficient Long-Horizon Agent Pruning via Trajectory-Anchored Recovery — arXiv:2610.09074

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09074
License (as archived)
CC0-1.0
Archived
2026-10-08T11:56:19.649Z
SHA-256 of the archived text
62a7c2b92c28dd51ca1bfc83eebb958b8519f4a07c212edf4ffb32066758e472
Stable URL of this page
https://metraipolis.com/documento/dc_1987eca5f458

How to cite

TAP: Efficient Long-Horizon Agent Pruning via Trajectory-Anchored Recovery — arXiv:2610.09074 — arXiv (en), https://arxiv.org/abs/2610.09074 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1987eca5f458 (Metraipolis archive dc_1987eca5f458, sha256 62a7c2b92c28dd51ca1bfc83eebb958b8519f4a07c212edf4ffb32066758e472)
BibTeX
@misc{dc1987eca5f458,
  title = {TAP: Efficient Long-Horizon Agent Pruning via Trajectory-Anchored Recovery — arXiv:2610.09074},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09074},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1987eca5f458; sha256 62a7c2b92c28dd51ca1bfc83eebb958b8519f4a07c212edf4ffb32066758e472},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09074. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Emerging long-horizon agentic tasks require repeated model calls, worsening the inference cost of already-costly language models. While narrow agentic tasks suggest potential for aggressive model pruning without performance drop, empirical results show existing methods proposed for question answering tasks severely degrade task performance when applied to agentic models. We trace this failure to two decisions: what to prune and how to recover. For pruning, one-shot importance estimates fail to track how the pruned model adapts. For recovery, offline distillation covers only teacher prefixes, while full-trajectory on-policy distillation causes student errors to compound across turns. In this work, we propose Trajectory-Anchored Pruning (TAP), the first structural pruning framework for reinforcement learning (RL)-trained agents. TAP couples structural pruning with efficient on-policy recovery, anchoring interactions to teacher trajectories while allowing the student to generate each reasoning-action response. A frozen dense teacher supervises the student's response prefixes, addressing within-response training-inference mismatch while preventing student-induced deviations from propagating across training turns. Instead of one-shot pruning, TAP re-scores channels using gradients of the recovery objective on the recovered student, connecting iterative channel selection to the evolving policy. With 60% of FFN channels removed, TAP retains 99.2% and 88.0% of the dense 7B agents' task success rates on ALFWorld and WebShop, respectively, while reducing GPU time per successful task by approximately 22% and 17%. These results demonstrate effective structural compression of long-horizon agents under a limited recovery budget.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.09074 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T10:50:37.651Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).