Apr 22
Talking to a Know-It-All GPT or a Second-Guesser Claude? How Repair reveals unreliable Multi-Turn Behavior in LLMs
★★★★★
significance 3/5
This study investigates how Large Language Models (LLMs) handle conversational 'repair' during multi-turn dialogues, specifically regarding math questions. The researchers found that different models exhibit varying levels of resistance or susceptibility to user-initiated corrections, leading to unpredictable behavior in long-form interactions.
Why it matters
Discrepancies in how models accept user corrections signal fundamental gaps in reasoning reliability and conversational control during complex, multi-turn interactions.
Tags
#llm reliability #multi-turn dialogue #conversational repair #error correctionRelated coverage
- Global South OpportunitiesPivotal Research Fellowship 2026 (Q3): AI Safety Research Opportunity - Global South Opportunities
- arXiv cs.AIAn Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge Enhancement
- arXiv cs.AIPExA: Parallel Exploration Agent for Complex Text-to-SQL
- arXiv cs.AIThe Power of Power Law: Asymmetry Enables Compositional Reasoning
- arXiv cs.AIOn the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation