arXiv：cs.CL· Obada Kraishan·· 3 小时前AI 评分53
研究：加大推理预算并不能减少大语言模型的认知偏差
The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models
AI 导读
arXiv 论文（2610.10049，已被 IEEE CogMI 2026 接收）通过 30 个情境、六类偏差、四个模型家族、12350 次 API 调用发现：推理模型并不比同源非推理模型更少偏差，实际消耗的推理 token 增加也不可靠地降低偏差幅度。
正文
Abstract: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. 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. Because a requested ceiling is not the same as realized deliberation, we use the reasoning tokens each call consumed as the dose. 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). Second, bias magnitude does not reliably fall as realized deliberation grows: no slope is significantly negative, and where anything moves it is the signed score drifting further from the human direction. Third, anchoring is the only bias in the human direction (d = 1.89). Four of the other five lean the opposite way in all seven models; with five items per bias, that reversal is reliable for framing and directional for escalation of commitment, confirmation, and loss aversion, while availability is absent. A one-line instruction to restate the anchor before answering lowered anchoring on all five anchoring items, which no amount of additional thinking did, although the effect does not reach significance (p = .057). The results argue against treating test-time reasoning as a rationality guarantee and for auditing deployed models bias by bias.
| Comments: | Accepted at the 2026 IEEE 8th International Conference on Cognitive Machine Intelligence (IEEE CogMI 2026). 8 pages, 3 figures, 5 tables. Code and data: this https URL |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.10049 [cs.CL] |
| (or arXiv:2610.10049v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10049 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Obada Kraishan [view email]
[v1]
Wed, 7 Oct 2026 13:23:51 UTC (74 KB)
来源：arXiv：cs.CL · arxiv.org