Metraipolis › archive › dc_b0426eca8a19

TraceSpec: Understanding and Improving Speculative Decoding for LLM Agents — arXiv:2610.41177

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41177
License (as archived)
CC0-1.0
Archived
2026-11-03T07:45:25.175Z
SHA-256 of the archived text
75f3de93f6d7e28faf0f6b33de2342819842cddf9f8255d82d9b1426ef0b3521
Stable URL of this page
https://metraipolis.com/documento/dc_b0426eca8a19

How to cite

TraceSpec: Understanding and Improving Speculative Decoding for LLM Agents — arXiv:2610.41177 — arXiv (en), https://arxiv.org/abs/2610.41177 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b0426eca8a19 (Metraipolis archive dc_b0426eca8a19, sha256 75f3de93f6d7e28faf0f6b33de2342819842cddf9f8255d82d9b1426ef0b3521)
BibTeX
@misc{dcb0426eca8a19,
  title = {TraceSpec: Understanding and Improving Speculative Decoding for LLM Agents — arXiv:2610.41177},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41177},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b0426eca8a19; sha256 75f3de93f6d7e28faf0f6b33de2342819842cddf9f8255d82d9b1426ef0b3521},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41177. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM agents execute tasks through repeated model calls over evolving histories of reasoning, actions, and tool feedback. Aggregate speculative-decoding metrics merge these distinct decisions, providing limited guidance on where draft supervision is needed. We present TraceSpec, a framework connecting trajectory evaluation with adaptive draft training. It combines dynamic execution with fixed-prefix replay and introduces TraceSpec-Traj, a corpus of 10,000 training states and 2,500 test states spanning browsing and software engineering. Analysis of five speculative baselines reveals that aggregate acceptance can rise with context length while declining within matched late browsing actions. Workload and phase further distinguish these trends, motivating supervision guided by trajectory conditions. TraceSpec trains on intermediate assistant turns and allocates a fixed prediction budget using state-conditioned estimates of prefix-aware acceptance gaps. Across three target sizes and two temperatures, it improves replay throughput by 12.5% to 22.6% over its DSpark base. Qwen3-4B ablations show that adaptive allocation increases accepted length by 3.3% to 10.0% over fixed full-turn supervision, with additional comparisons substantiating the roles of state conditioning and consecutive-token agreement. The resulting evaluation resources, trajectory findings, and training strategy connect the analysis of agent execution to practical draft improvement.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.41177 — 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 2026-11-03T07:39:52.166Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).