Metraipolis › archive › dc_12bf626d1d8a

ActionOPD: Distilling Decisions, Not Tokens — arXiv:2610.19820

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19820
License (as archived)
CC0-1.0
Archived
2026-10-15T20:02:05.117Z
SHA-256 of the archived text
e53dd0999701c94c61a2cef8accdbe9c2595b59796d217212c12f79dc5e7f182
Stable URL of this page
https://metraipolis.com/documento/dc_12bf626d1d8a

How to cite

ActionOPD: Distilling Decisions, Not Tokens — arXiv:2610.19820 — arXiv (en), https://arxiv.org/abs/2610.19820 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_12bf626d1d8a (Metraipolis archive dc_12bf626d1d8a, sha256 e53dd0999701c94c61a2cef8accdbe9c2595b59796d217212c12f79dc5e7f182)
BibTeX
@misc{dc12bf626d1d8a,
  title = {ActionOPD: Distilling Decisions, Not Tokens — arXiv:2610.19820},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19820},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_12bf626d1d8a; sha256 e53dd0999701c94c61a2cef8accdbe9c2595b59796d217212c12f79dc5e7f182},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19820. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Distilling capable language-model agents into smaller models requires aligning the student's behavior with the teacher's. This raises a fundamental question: at which states and at what granularity should this alignment be performed? Offline imitation provides dense but token-specific supervision and can suffer from state-distribution mismatch. Token-level on-policy distillation (OPD) addresses this mismatch but typically requires teacher token probabilities and compatible tokenization. We argue that semantic actions provide a natural unit of supervision for agent distillation. We introduce ActionOPD, an action-level on-policy distillation method motivated by maximizing expected teacher-action likelihood at student-encountered states. At each such state, ActionOPD samples actions from both the student and a black-box teacher, rewarding student actions according to their semantic agreement with the sampled teacher actions. Through policy optimization, ActionOPD aligns the student's action-level behavior with the teacher's decisions without requiring reproduction of teacher-written reasoning or access to teacher token probabilities. Training Qwen3-4B and Qwen3-8B with two teachers on diverse, curated tool-use tasks yields average gains of 3.3-6.9 percentage points across four out-of-distribution benchmarks. ActionOPD scales favorably with student size and teacher capability, without requiring the teacher and student to share a model family or tokenizer. On $\tau^2$-bench, ActionOPD achieves competitive performance against offline distillation and token-level OPD. We further explore Self-ActionOPD with a same-model teacher and privileged context. These results establish ActionOPD as a promising approach to agent distillation.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.19820 — 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 2026-10-15T19:35:48.270Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).