Metraipolis › archive › dc_1daee7484d41

Agentic Scaffolding Amplifies Sycophantic Behavior in Large Language Models — arXiv:2608.21377

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.21377
License (as archived)
CC0-1.0
Archived
2026-10-18T07:42:08.426Z
SHA-256 of the archived text
e28c148cbd969303990f9c42a2eb2d041652b785097bd644710a3e4956155c52
Stable URL of this page
https://metraipolis.com/documento/dc_1daee7484d41

How to cite

Agentic Scaffolding Amplifies Sycophantic Behavior in Large Language Models — arXiv:2608.21377 — arXiv (en), https://arxiv.org/abs/2608.21377 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1daee7484d41 (Metraipolis archive dc_1daee7484d41, sha256 e28c148cbd969303990f9c42a2eb2d041652b785097bd644710a3e4956155c52)
BibTeX
@misc{dc1daee7484d41,
  title = {Agentic Scaffolding Amplifies Sycophantic Behavior in Large Language Models — arXiv:2608.21377},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.21377},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1daee7484d41; sha256 e28c148cbd969303990f9c42a2eb2d041652b785097bd644710a3e4956155c52},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.21377. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Sycophancy in large language models, the tendency to prioritize user agreement over truthful responses, has been documented extensively but studied primarily in single-turn settings. This paper investigates a critical question: does subjecting LLMs to greater interaction scaffolding make sycophancy better or worse? Across 4,800 veracity judgments (200 statements $\times$ 6 models $\times$ 4 conditions), we find that the interaction scaffolding characteristic of agentic systems (feedback loops, reconsideration checkpoints, and iterative refinement) systematically amplifies sycophantic behavior. Multi-turn interaction, user pressure, and iterative self-refinement each provide additional opportunities for models to drift toward agreement, and this drift coincides with a mean accuracy drop of $-6.3$ percentage points, establishing the capitulation as harmful rather than corrective. More capable models show larger amplification effects, a troubling inversion of expectations. We introduce the concept of agentic sycophancy amplification (ASA) and two novel metrics: capitulation rate and sycophantic capitulation rate. Our results indicate that as AI systems acquire greater autonomy, sycophancy becomes compounding rather than merely persistent. Systems designed with human oversight loops may inadvertently create the conditions for this drift.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2608.21377 — 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 2026-10-18T07:37:52.795Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).