Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection

• URL: http://arxiv.org/abs/2510.21049v1
• Date: Thu, 23 Oct 2025 23:23:36 GMT
• Title: Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection
• Authors: Atoosa Chegini, Hamid Kazemi, Garrett Souza, Maria Safi, Yang Song, Samy Bengio, Sinead Williamson, Mehrdad Farajtabar,
• Abstract summary: Reasoning has become a central paradigm for large language models (LLMs)<n>We present the first systematic study of reasoning for classification tasks under strict low false positive rate regimes.<n>Our results reveal a clear trade-off: Think On (reasoning-augmented) generation improves overall accuracy, but underperforms at the low-FPR thresholds essential for practical use.
• Score: 21.190105743961798
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reasoning has become a central paradigm for large language models (LLMs), consistently boosting accuracy across diverse benchmarks. Yet its suitability for precision-sensitive tasks remains unclear. We present the first systematic study of reasoning for classification tasks under strict low false positive rate (FPR) regimes. Our analysis covers two tasks--safety detection and hallucination detection--evaluated in both fine-tuned and zero-shot settings, using standard LLMs and Large Reasoning Models (LRMs). Our results reveal a clear trade-off: Think On (reasoning-augmented) generation improves overall accuracy, but underperforms at the low-FPR thresholds essential for practical use. In contrast, Think Off (no reasoning during inference) dominates in these precision-sensitive regimes, with Think On surpassing only when higher FPRs are acceptable. In addition, we find token-based scoring substantially outperforms self-verbalized confidence for precision-sensitive deployments. Finally, a simple ensemble of the two modes recovers the strengths of each. Taken together, our findings position reasoning as a double-edged tool: beneficial for average accuracy, but often ill-suited for applications requiring strict precision.

Related papers

• Thinking by Subtraction: Confidence-Driven Contrastive Decoding for LLM Reasoning [58.331709210563616]
  Thinking by Subtraction is a confidence-driven contrastive decoding approach.<n>A small subset of low-confidence tokens disproportionately contributes to reasoning errors and unnecessary output expansion.<n>Our method, Confidence-Driven Contrastive Decoding, detects low-confidence tokens during decoding and intervenes at these positions.
  arXiv  Detail & Related papers  (2026-02-20T14:13:22Z)
• Decomposed Prompting Does Not Fix Knowledge Gaps, But Helps Models Say "I Don't Know" [47.930782177987446]
  Large language models often struggle to recognize their knowledge limits in closed-book question answering, leading to confident hallucinations.<n>We evaluate three task-equivalent prompting regimes: Direct, Assistive, and Incremental, across different model scales and multi-hop QA benchmarks.<n>Because factual knowledge is stable while hallucinations are agreement, cross-regime provides a precise signal of internal uncertainty.
  arXiv  Detail & Related papers  (2026-02-04T18:39:58Z)
• Outcome Accuracy is Not Enough: Aligning the Reasoning Process of Reward Models [108.26461635308796]
  We introduce Rationale Consistency, a fine-grained metric that quantifies the alignment between the model's reasoning process and human judgment.<n>Our evaluation of frontier models reveals that rationale consistency effectively discriminates among state-of-the-art models.<n>We introduce a hybrid signal that combines rationale consistency with outcome accuracy for GenRM training.
  arXiv  Detail & Related papers  (2026-02-04T15:24:52Z)
• A Comprehensive Evaluation of LLM Reasoning: From Single-Model to Multi-Agent Paradigms [20.241519889633285]
  Large Language Models (LLMs) are increasingly deployed as reasoning systems, where reasoning paradigms play a critical role.<n>We conduct a comprehensive and unified evaluation of reasoning paradigms, spanning direct single-model generation, CoT-augmented single-model reasoning, and representative MAS.<n>We introduce MIMeBench, a new open-ended benchmark that targets two foundational yet underexplored semantic capabilities.
  arXiv  Detail & Related papers  (2026-01-19T17:23:45Z)
• Mitigating LLM Hallucination via Behaviorally Calibrated Reinforcement Learning [32.32593439144886]
  Behavior-calibrated reinforcement learning allows smaller models to surpass frontier models in uncertainty quantification.<n>Our model's log-scale Accuracy-to-Hallucination Ratio gain (0.806) exceeds GPT-5's (0.207) in a challenging in-domain evaluation.
  arXiv  Detail & Related papers  (2025-12-22T22:51:48Z)
• Rubric-Conditioned LLM Grading: Alignment, Uncertainty, and Robustness [4.129847064263056]
  We systematically evaluate the performance of Large Language Models for rubric-based short-answer grading.<n>We find that alignment is strong for binary tasks but degrades with increased rubric granularity.<n>Experiments reveal that while the model is resilient to prompt injection, it is sensitive to synonym substitutions.
  arXiv  Detail & Related papers  (2025-12-21T05:22:04Z)
• Do LLMs Know They Are Being Tested? Evaluation Awareness and Incentive-Sensitive Failures in GPT-OSS-20B [1.948261185683419]
  We investigate whether "evaluation scent" inflates measured performance without commensurate capability gains.<n>We run six paired A/B scenarios that hold task content and decoding fixed while varying framing.<n>We provide a reproducible A/B framework (prompt banks, validators, per-run scores, scripts) and practical guidance.
  arXiv  Detail & Related papers  (2025-10-08T09:49:05Z)
• Judging with Confidence: Calibrating Autoraters to Preference Distributions [56.17041629492863]
  We argue that a reliable autorater must learn to model the full distribution of preferences defined by a target population.<n>We present two learning methods tailored to different data conditions.<n>Our results show that finetuning autoraters with a distribution-matching objective leads to verbalized probability predictions that are better aligned with the target preference distribution.
  arXiv  Detail & Related papers  (2025-09-30T20:36:41Z)
• Inducing Faithfulness in Structured Reasoning via Counterfactual Sensitivity [6.908972852063454]
  Large language models often generate a correct answer while relying on a flawed or irrelevant reasoning trace.<n>This paper introduces textbfCounterfactual Sensitivity Regularization (CSR), a novel training objective.<n>CSR improves faithfulness over standard fine-tuning and process supervision by up to 70 percentage points.
  arXiv  Detail & Related papers  (2025-09-01T15:18:46Z)
• Beyond Binary Rewards: Training LMs to Reason About Their Uncertainty [59.97939500426759]
  This paper describes RLCR, an approach to training reasoning models that jointly improves accuracy and confidence estimation.<n>We show that across diverse datasets, RLCR substantially improves calibration with no loss in accuracy.<n>We also demonstrate that verbalized confidence can be leveraged at test time to improve accuracy and calibration.
  arXiv  Detail & Related papers  (2025-07-22T17:56:01Z)
• Towards Evaluting Fake Reasoning Bias in Language Models [47.482898076525494]
  We show that models favor the surface structure of reasoning even when the logic is flawed.<n>We introduce THEATER, a benchmark that systematically investigates Fake Reasoning Bias (FRB)<n>We evaluate 17 advanced Large Language Models (LRMs) on both subjective DPO and factual datasets.
  arXiv  Detail & Related papers  (2025-07-18T09:06:10Z)
• Mind the Confidence Gap: Overconfidence, Calibration, and Distractor Effects in Large Language Models [0.6091702876917281]
  Large Language Models (LLMs) show remarkable proficiency in natural language tasks.<n>Overconfidence-misalignment between predicted confidence and true correctness poses significant risks in critical decision-making applications.<n>We present a comprehensive analysis on calibration in LLMs across nine LLMs and three factual Question-Answering datasets.
  arXiv  Detail & Related papers  (2025-02-16T07:46:09Z)
• Evaluating Human Alignment and Model Faithfulness of LLM Rationale [66.75309523854476]
  We study how well large language models (LLMs) explain their generations through rationales. We show that prompting-based methods are less "faithful" than attribution-based explanations.
  arXiv  Detail & Related papers  (2024-06-28T20:06:30Z)
• Are We Really Achieving Better Beyond-Accuracy Performance in Next Basket Recommendation? [57.91114305844153]
  Next basket recommendation (NBR) is a special type of sequential recommendation that is increasingly receiving attention. Recent studies into NBR have found a substantial performance difference between recommending repeat items and explore items. We propose a plug-and-play two-step repetition-exploration framework that treats repeat items and explores items separately.
  arXiv  Detail & Related papers  (2024-05-02T09:59:35Z)
• Worst Case Matters for Few-Shot Recognition [27.023352955311502]
  Few-shot recognition learns a recognition model with very few (e.g., 1 or 5) images per category. Current few-shot learning methods focus on improving the average accuracy over many episodes. We argue that in real-world applications we may often only try one episode instead of many, and hence maximizing the worst-case accuracy is more important than maximizing the average accuracy.
  arXiv  Detail & Related papers  (2022-03-13T05:39:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.