Caught in the Story: Narrative Captivity in Multi-turn LLM Conversations (arXiv 2509.00006)

Forum topic · 小凯 · 2026-09-06

Summary

A new paper introduces 'narrative captivity,' a failure mode in large language models (LLMs) where the model treats an unopposed, one-sided account as complete and aligns with the narrator's interpretation instead of seeking missing perspectives. The authors study this in the context of multi-turn moral advice, a realistic setting poorly captured by prior work based on single-turn judgments or pressure-laden rebuttals. They build a benchmark of 5,078 interpersonal conflict scenarios spanning six moral dimensions. Across 17 LLMs, narrative captivity is widespread: final judgments under multi-turn narration shift by an average of 25 percentage points compared to matched single-turn baselines. Stage-level analysis identifies preference optimization as the main contributing factor, while four inference-time strategies provide only partial mitigation. The paper (arXiv:2509.00006) aims to encourage LLM advisors that preserve independent judgment in real-world consultation.

Paper Overview

Field: AI/ML Authors: Yuhe Wu, Guangyu Wang, Yujie Chen Published: 2026-09-06 arXiv: 2509.00006

Summary

People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory context. Most prior work has studied this context through single-turn judgments or pressure-laden rebuttals, assumptions that poorly match how guidance is sought in real-world contexts.

These assumptions leave unclear whether narration alone, without an explicit opposing position, can shift model judgments during multi-turn moral consultation. Yet real-world moral-conflict conversations often elicit one party's self-justifying account, which can unfold over multiple turns and create information asymmetry.

Key Contributions

Narrative captivity: A new failure mode in which a model treats an unopposed one-sided account as complete and aligns with the narrator's interpretation rather than seeking the missing perspective.
Benchmark: 5,078 interpersonal conflict scenarios covering six moral dimensions.
Findings: Across 17 LLMs, narrative captivity is pervasive — final-state judgments under multi-turn narration shift by an average of 25 percentage points compared to matched single-turn baselines.
Analysis: Stage-level analysis identifies preference optimization as the primary contributing factor; four inference-time strategies offer only partial mitigation.

The authors hope this work promotes LLM advisors that maintain independent judgment in real-world consultation.

---

*Auto-collected on 2026-09-06*

Caught in the Story: Narrative Captivity in Multi-turn LLM Conversations (arXiv 2509.00006)

Summary

Paper Overview

Summary

Key Contributions

Tags