Metraipolis › archive › dc_df74bdf98959

TAC-CoT: Tool-Empowered Implicit Chain-of-Thought via Interleaving Explicit Tokens — arXiv:2611.01926

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.01926
License (as archived): CC0-1.0
Archived: 2026-11-03T11:10:39.138Z
SHA-256 of the archived text: df517bb6d393ebd2d2b675569dababdf688dd56aae96e8135fa30abde830abfb
Stable URL of this page: https://metraipolis.com/documento/dc_df74bdf98959

How to cite

TAC-CoT: Tool-Empowered Implicit Chain-of-Thought via Interleaving Explicit Tokens — arXiv:2611.01926 — arXiv (en), https://arxiv.org/abs/2611.01926 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_df74bdf98959 (Metraipolis archive dc_df74bdf98959, sha256 df517bb6d393ebd2d2b675569dababdf688dd56aae96e8135fa30abde830abfb)

BibTeX

@misc{dcdf74bdf98959,
  title = {TAC-CoT: Tool-Empowered Implicit Chain-of-Thought via Interleaving Explicit Tokens — arXiv:2611.01926},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01926},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_df74bdf98959; sha256 df517bb6d393ebd2d2b675569dababdf688dd56aae96e8135fa30abde830abfb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01926. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM-based agentic systems are increasingly combined with CoT reasoning to solve tasks that require accurate reasoning, external knowledge, and multi-step deliberation. Yet, explicit CoT incurs substantial generation overhead, while implicit CoT, which replaces textual rationales with compact continuous tokens, does not naturally support the explicit communication interface required for agentic tool-use. We propose TAC-CoT, a tool-empowered adaptive implicit CoT framework that replaces explicit reasoning traces with compact latent reasoning segments while retaining textual tokens only for tool calls, observations, and final answers. This interleaved design enables agentic tool-use without verbalizing every intermediate reasoning step. Experiments on agentic math and QA benchmarks show that TAC-CoT improves accuracy by up to 8.8 points while reducing model forward passes by up to 6.8x. Furthermore, analyses show that TAC-CoT improves tool-use quality in both single- and multi-tool settings.

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/2611.01926 — 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 2026-11-03T10:59:52.317Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).