Metraipolis › archive › dc_405cf71897ea

Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise — arXiv:2610.02702

Source: arXiv
Original: https://arxiv.org/abs/2610.02702
License (as archived): CC0-1.0
Archived: 2026-10-05T22:48:04.743Z
SHA-256 of the archived text: 94f38316b966564b84fbb80716818b515db6fbf9a3bd444ecfec84543865aebf
Stable URL of this page: https://metraipolis.com/documento/dc_405cf71897ea

How to cite

Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise — arXiv:2610.02702 — arXiv, https://arxiv.org/abs/2610.02702 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_405cf71897ea (Metraipolis archive dc_405cf71897ea, sha256 94f38316b966564b84fbb80716818b515db6fbf9a3bd444ecfec84543865aebf)

BibTeX

@misc{dc405cf71897ea,
  title = {Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise — arXiv:2610.02702},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.02702},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_405cf71897ea; sha256 94f38316b966564b84fbb80716818b515db6fbf9a3bd444ecfec84543865aebf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02702. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multi-agent debate is increasingly used to reach consensus among LLM agents, yet agents often yield to a unanimous majority. When an agent changes its answer, has it changed its mind or only its statement? We study this with two-hop factual questions whose intermediate entity (the bridge, e.g. the country in "the capital of the country where the Sagrada Familia is located") is never stated by anyone. Scripted peers, in the role of Asch's confederates, unanimously assert a wrong answer taken from another fact with a different bridge. At the moment the agent answers, we read the bridge from its residual stream with the Jacobian lens (J-lens) and, for comparison, the logit lens. In pre-registered tests on held-out facts with four open-weight models, agents of Qwen3.5-4B, Qwen3.6-27B and Gemma-4-E4B-it that gave in still represented their original bridge in the pre-registered layers below the output (hit@100 above a control entity: 0.85, 0.22 and 0.24), where the logit lens rarely ranked it among the top 100 tokens (0.00-0.06). These agents also represented the bridge behind the peers' answer, beyond a mention baseline. A pre-registered addendum hid the agent's earlier answer or removed it: agents that gave in still represented their original bridge in all four models (0.43, 0.29, 0.37 and 0.25 with the answer hidden), including Llama-3.1-8B-Instruct, which barely did so with its answer in view (0.03). The premise can thus be computed from the question alone while the agent states the majority's answer. Hiding the earlier answer also changed conformity: Qwen3.5-4B gave in on 89% of questions instead of 8%. In exploratory interventions, injecting the bridge's J-lens direction brought agents back to their original answer only in the two Qwen models. Stated consensus in multi-agent debate can thus overstate agreement. We also report the negative results of our pre-registered program.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2610.02702 — 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 2026-10-05T22:45:40.645Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).