Paper
Thinking to Recall: How Reasoning Unlocks Parametric Knowledge in LLMs
Authors
Zorik Gekhman, Roee Aharoni, Eran Ofek, Mor Geva, Roi Reichart, Jonathan Herzig
Abstract
While reasoning in LLMs plays a natural role in math, code generation, and multi-hop factual questions, its effect on simple, single-hop factual questions remains unclear. Such questions do not require step-by-step logical decomposition, making the utility of reasoning highly counterintuitive. Nevertheless, we find that enabling reasoning substantially expands the capability boundary of the model's parametric knowledge recall, unlocking correct answers that are otherwise effectively unreachable. Why does reasoning aid parametric knowledge recall when there are no complex reasoning steps to be done? To answer this, we design a series of hypothesis-driven controlled experiments, and identify two key driving mechanisms: (1) a computational buffer effect, where the model uses the generated reasoning tokens to perform latent computation independent of their semantic content; and (2) factual priming, where generating topically related facts acts as a semantic bridge that facilitates correct answer retrieval. Importantly, this latter generative self-retrieval mechanism carries inherent risks: we demonstrate that hallucinating intermediate facts during reasoning increases the likelihood of hallucinations in the final answer. Finally, we show that our insights can be harnessed to directly improve model accuracy by prioritizing reasoning trajectories that contain hallucination-free factual statements.
Metadata
Related papers
Gen-Searcher: Reinforcing Agentic Search for Image Generation
Kaituo Feng, Manyuan Zhang, Shuang Chen, Yunlong Lin, Kaixuan Fan, Yilei Jian... • 2026-03-30
On-the-fly Repulsion in the Contextual Space for Rich Diversity in Diffusion Transformers
Omer Dahary, Benaya Koren, Daniel Garibi, Daniel Cohen-Or • 2026-03-30
Graphilosophy: Graph-Based Digital Humanities Computing with The Four Books
Minh-Thu Do, Quynh-Chau Le-Tran, Duc-Duy Nguyen-Mai, Thien-Trang Nguyen, Khan... • 2026-03-30
ParaSpeechCLAP: A Dual-Encoder Speech-Text Model for Rich Stylistic Language-Audio Pretraining
Anuj Diwan, Eunsol Choi, David Harwath • 2026-03-30
RAD-AI: Rethinking Architecture Documentation for AI-Augmented Ecosystems
Oliver Aleksander Larsen, Mahyar T. Moghaddam • 2026-03-30
Raw Data (Debug)
{
"raw_xml": "<entry>\n <id>http://arxiv.org/abs/2603.09906v1</id>\n <title>Thinking to Recall: How Reasoning Unlocks Parametric Knowledge in LLMs</title>\n <updated>2026-03-10T16:59:20Z</updated>\n <link href='https://arxiv.org/abs/2603.09906v1' rel='alternate' type='text/html'/>\n <link href='https://arxiv.org/pdf/2603.09906v1' rel='related' title='pdf' type='application/pdf'/>\n <summary>While reasoning in LLMs plays a natural role in math, code generation, and multi-hop factual questions, its effect on simple, single-hop factual questions remains unclear. Such questions do not require step-by-step logical decomposition, making the utility of reasoning highly counterintuitive. Nevertheless, we find that enabling reasoning substantially expands the capability boundary of the model's parametric knowledge recall, unlocking correct answers that are otherwise effectively unreachable. Why does reasoning aid parametric knowledge recall when there are no complex reasoning steps to be done? To answer this, we design a series of hypothesis-driven controlled experiments, and identify two key driving mechanisms: (1) a computational buffer effect, where the model uses the generated reasoning tokens to perform latent computation independent of their semantic content; and (2) factual priming, where generating topically related facts acts as a semantic bridge that facilitates correct answer retrieval. Importantly, this latter generative self-retrieval mechanism carries inherent risks: we demonstrate that hallucinating intermediate facts during reasoning increases the likelihood of hallucinations in the final answer. Finally, we show that our insights can be harnessed to directly improve model accuracy by prioritizing reasoning trajectories that contain hallucination-free factual statements.</summary>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n <published>2026-03-10T16:59:20Z</published>\n <arxiv:primary_category term='cs.CL'/>\n <author>\n <name>Zorik Gekhman</name>\n </author>\n <author>\n <name>Roee Aharoni</name>\n </author>\n <author>\n <name>Eran Ofek</name>\n </author>\n <author>\n <name>Mor Geva</name>\n </author>\n <author>\n <name>Roi Reichart</name>\n </author>\n <author>\n <name>Jonathan Herzig</name>\n </author>\n </entry>"
}