# Failing to Falsify: Evaluating and Mitigating Confirmation Bias in Language Models

arXiv: https://arxiv.org/abs/2604.02485 · HTML: https://mlsanity.com/paper/2604.02485/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 19 of its 33 references that are in the map.
- First posted on arXiv 2 Apr 2026.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 19 of 19)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [A Survey on Large Language Model based Autonomous Agents](https://mlsanity.com/paper/2308.11432/) (2023)
- [AlphaEvolve: A coding agent for scientific and algorithmic discovery](https://mlsanity.com/paper/2506.13131/) (2025)
- [The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search](https://mlsanity.com/paper/2504.08066/) (2025)
- [Symbolic Knowledge Distillation: from General Language Models to Commonsense Models](https://mlsanity.com/paper/2110.07178/) (2021)
- [Language models show human-like content effects on reasoning tasks](https://mlsanity.com/paper/2207.07051/) (2022)
- [Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worse](https://mlsanity.com/paper/2410.21333/) (2024)
- [An AI system to help scientists write expert-level empirical software](https://mlsanity.com/paper/2509.06503/) (2025)
- [Large Language Models Assume People are More Rational than We Really are](https://mlsanity.com/paper/2406.17055/) (2024)
- [A Comprehensive Evaluation of Cognitive Biases in LLMs](https://mlsanity.com/paper/2410.15413/) (2024)
- [Doing Experiments and Revising Rules with Natural Language and Probabilistic Reasoning](https://mlsanity.com/paper/2402.06025/) (2024)
- [AI Can Be Cognitively Biased: An Exploratory Study on Threshold Priming in LLM-Based Batch Relevance Assessment](https://mlsanity.com/paper/2409.16022/) (2024)
- [What is a Number, That a Large Language Model May Know It?](https://mlsanity.com/paper/2502.01540/) (2025)
- [Unveiling Confirmation Bias in Chain-of-Thought Reasoning](https://mlsanity.com/paper/2506.12301/) (2025)
- [WILT: A Multi-Turn, Memorization-Robust Inductive Logic Benchmark for LLMs](https://mlsanity.com/paper/2410.10998/) (2024)
- [Language Agents Mirror Human Causal Reasoning Biases. How Can We Help Them Think Like Scientists?](https://mlsanity.com/paper/2505.09614/) (2025)
- [A Rational Analysis of the Speech-to-Song Illusion](https://mlsanity.com/paper/2402.06992/) (2024)

## Built on by (most cited first, 9 of 9)

- [When Stale Constraints Go Unchecked: Budgeted Verification Failures in Inherited Agent Memory](https://mlsanity.com/paper/2608.25553/) (2026)
- [FALSIFYBENCH: Evaluating Inductive Reasoning in LLMs with Rule Discovery Games](https://mlsanity.com/paper/2606.04751/) (2026)
- [CogEEGAgent: Toward Autonomous Cognitive EEG Analysis with Grounded Execution and Selection-Aware Verification](https://mlsanity.com/paper/2607.25045/) (2026)
- [Diagnostic Foundation for Evaluating LLMs' Research Integrity as Co-Scientists](https://mlsanity.com/paper/2608.12345/) (2026)
- [Plan Pointers and Record-Directive Form in Budgeted Verification of Inherited Agent Memory](https://mlsanity.com/paper/2609.03450/) (2026)
- [Witness: Discovery, Deciphering, and Epiphany in Interactive Puzzle Environments](https://mlsanity.com/paper/2609.32208/) (2026)
- ["You're Right, Let Me Fix It": How LLM Agents Damage Correct Work When Falsely Accused](https://mlsanity.com/paper/2609.32616/) (2026)
- [Trust the Brand, Lose Control: How Identity Hijacks LLM Agent Orchestration](https://mlsanity.com/paper/2609.32635/) (2026)
- [DoGBench: Can Agents Meet Expert Standards for User-Facing Documentation?](https://mlsanity.com/paper/2609.39909/) (2026)

## Abstract (arXiv)

Confirmation bias, the tendency to seek evidence that supports rather than challenges one's belief, hinders one's reasoning ability. We examine whether large language models (LLMs) exhibit confirmation bias by adapting the rule-discovery study from human psychology: given a sequence of three numbers (a "triple"), an agent engages in an interactive feedback loop where it (1) proposes a new triple, (2) receives feedback on whether it satisfies the hidden rule, and (3) guesses the rule. Across eleven LLMs of multiple families and scales, we find that LLMs exhibit confirmation bias, often proposing triples to confirm their hypothesis rather than trying to falsify it. This leads to slower and less frequent discovery of the hidden rule. We further explore intervention strategies (e.g., encouraging the agent to consider counter examples) developed for humans. We find prompting LLMs with such instruction consistently decreases confirmation bias in LLMs, improving rule discovery rates from 42% to 56% on average. Lastly, we mitigate confirmation bias by distilling intervention-induced behavior into LLMs, showing promising generalization to a new task, the Blicket test. Our work shows that confirmation bias is a limitation of LLMs in hypothesis exploration, and that it can be mitigated via injecting interventions designed for humans.