Metraipolis › archive › dc_7cae0ffc50db

You Can't Escape Your Own Activations : Evaluation Awareness and Multi-Agent Monitoring — arXiv:2609.03035

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.03035
License (as archived): CC0-1.0
Archived: 2026-10-17T01:02:58.288Z
SHA-256 of the archived text: 3ae33498287b1611abd8e1eda0ea90f776bf48c81afbe8c645bb39a40a056c31
Stable URL of this page: https://metraipolis.com/documento/dc_7cae0ffc50db

How to cite

You Can't Escape Your Own Activations : Evaluation Awareness and Multi-Agent Monitoring — arXiv:2609.03035 — arXiv (en), https://arxiv.org/abs/2609.03035 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7cae0ffc50db (Metraipolis archive dc_7cae0ffc50db, sha256 3ae33498287b1611abd8e1eda0ea90f776bf48c81afbe8c645bb39a40a056c31)

BibTeX

@misc{dc7cae0ffc50db,
  title = {You Can't Escape Your Own Activations : Evaluation Awareness and Multi-Agent Monitoring — arXiv:2609.03035},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.03035},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7cae0ffc50db; sha256 3ae33498287b1611abd8e1eda0ea90f776bf48c81afbe8c645bb39a40a056c31},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.03035. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM agents are increasingly deployed in multi-agent systems, where they can collude while keeping their actions benign. Output monitors designed to detect such collusions can be fooled by obfuscation and steganography, motivating the use of probes trained on internal activations. However, these probes are usually evaluated on agents that do not know they are being watched. We study how activation-based detection changes when agents are explicitly informed that their internal activations are being monitored, and when they additionally receive feedback from the monitor. We keep the models, probes, and thresholds fixed and change only what the agents are told: nothing (baseline), that an activation monitor is present (aware), or that a monitor is present together with the previous round's score (feedback). We test two games, a four-agent blackjack game and a two-agent Simmons prisoners game, using Qwen3-32B-AWQ and GPT-OSS-20B in homogeneous and mixed pairings. Telling agents about the monitor does not hide them. The best probes stay accurate in all three conditions, and the agents keep colluding.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.MA).
Categorie arXiv: Multiagent Systems (cs.MA) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.03035 — 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 2026-10-17T00:22:10.222Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).