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

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chegini, Atoosa, Kazemi, Hamid, Souza, Garrett, Safi, Maria, Song, Yang, Bengio, Samy, Williamson, Sinead, Farajtabar, Mehrdad
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912668203352064
author Chegini, Atoosa
Kazemi, Hamid
Souza, Garrett
Safi, Maria
Song, Yang
Bengio, Samy
Williamson, Sinead
Farajtabar, Mehrdad
author_facet Chegini, Atoosa
Kazemi, Hamid
Souza, Garrett
Safi, Maria
Song, Yang
Bengio, Samy
Williamson, Sinead
Farajtabar, Mehrdad
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2510_21049
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection
Chegini, Atoosa
Kazemi, Hamid
Souza, Garrett
Safi, Maria
Song, Yang
Bengio, Samy
Williamson, Sinead
Farajtabar, Mehrdad
Computation and Language
Artificial Intelligence
Machine Learning
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.
title Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.21049