The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models

Key points

• If this deliberation works the way dual-process accounts of human cognition suggest, longer thinking should weaken the classic decision biases that fast, intuitive judgment produces.1
• Using 30 vignettes covering six biases (anchoring, framing, loss aversion, escalation of commitment, availability, confirmation) from an established benchmark, we run a dose-response study across four model families, pairing each reasoning model with a matched non-reasoning sibling and requesting thinking ceilings of 0, 1,024, 4,096, and 8,192 tokens, for 12,350 API calls.1
• First, reasoning models are not less biased than their siblings; the point estimate leans the other way in every family, but the item-level pooled contrast is not reliable (Delta = +0.031, t(29) = 1.45, p = .157).1
• The results argue against treating test-time reasoning as a rationality guarantee and for auditing deployed models bias by bias.1

Sources (1)

• [1]The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 01:23 PM

  Reasoning models allocate extra computation at inference time and present their answers as the product of deliberate thought.

  If this deliberation works the way dual-process accounts of human cognition suggest, longer thinking should weaken the classic decision biases that fast, intuitive judgment produces.

Extractive summary: sentences quoted from the sources.