Beyond the Last Answer: Your Reasoning Trace Uncovers More than You Think
- URL: http://arxiv.org/abs/2504.20708v1
- Date: Tue, 29 Apr 2025 12:39:07 GMT
- Title: Beyond the Last Answer: Your Reasoning Trace Uncovers More than You Think
- Authors: Hasan Abed Al Kader Hammoud, Hani Itani, Bernard Ghanem,
- Abstract summary: We analyze intermediate reasoning steps, termed subthoughts, to answer two questions: Does the final answer reliably represent the model's optimal conclusion?<n>Our approach involves segmenting a reasoning trace into sequential subthoughts based on linguistic cues.<n>We find that aggregating these answers by selecting the most frequent one (the mode) often yields significantly higher accuracy compared to relying solely on the answer derived from the original complete trace.
- Score: 51.0691253204425
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Large Language Models (LLMs) leverage step-by-step reasoning to solve complex problems. Standard evaluation practice involves generating a complete reasoning trace and assessing the correctness of the final answer presented at its conclusion. In this paper, we challenge the reliance on the final answer by posing the following two questions: Does the final answer reliably represent the model's optimal conclusion? Can alternative reasoning paths yield different results? To answer these questions, we analyze intermediate reasoning steps, termed subthoughts, and propose a method based on our findings. Our approach involves segmenting a reasoning trace into sequential subthoughts based on linguistic cues. We start by prompting the model to generate continuations from the end-point of each intermediate subthought. We extract a potential answer from every completed continuation originating from different subthoughts. We find that aggregating these answers by selecting the most frequent one (the mode) often yields significantly higher accuracy compared to relying solely on the answer derived from the original complete trace. Analyzing the consistency among the answers derived from different subthoughts reveals characteristics that correlate with the model's confidence and correctness, suggesting potential for identifying less reliable answers. Our experiments across various LLMs and challenging mathematical reasoning datasets (AIME2024 and AIME2025) show consistent accuracy improvements, with gains reaching up to 13\% and 10\% respectively. Implementation is available at: https://github.com/hammoudhasan/SubthoughtReasoner.
Related papers
- Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification [23.190823296729732]
We study whether reasoning models encode information about answer correctness through probing the model's hidden states.<n>The resulting probe can verify intermediate answers with high accuracy and produces highly calibrated scores.
arXiv Detail & Related papers (2025-04-07T18:42:01Z) - Chain-of-Probe: Examining the Necessity and Accuracy of CoT Step-by-Step [81.50681925980135]
We propose a method to probe changes in the mind during the model's reasoning.<n>By analyzing patterns in mind change, we examine the correctness of the model's reasoning.<n>Our validation reveals that many responses, although correct in their final answer, contain errors in their reasoning process.
arXiv Detail & Related papers (2024-06-23T15:50:22Z) - Getting MoRE out of Mixture of Language Model Reasoning Experts [71.61176122960464]
We propose a Mixture-of-Reasoning-Experts (MoRE) framework that ensembles diverse specialized language models.
We specialize the backbone language model with prompts optimized for different reasoning categories, including factual, multihop, mathematical, and commonsense reasoning.
Our human study confirms that presenting expert predictions and the answer selection process helps annotators more accurately calibrate when to trust the system's output.
arXiv Detail & Related papers (2023-05-24T02:00:51Z) - ReCEval: Evaluating Reasoning Chains via Correctness and Informativeness [67.49087159888298]
ReCEval is a framework that evaluates reasoning chains via two key properties: correctness and informativeness.
We show that ReCEval effectively identifies various error types and yields notable improvements compared to prior methods.
arXiv Detail & Related papers (2023-04-21T02:19:06Z) - Measuring and Narrowing the Compositionality Gap in Language Models [116.5228850227024]
We measure how often models can correctly answer all sub-problems but not generate the overall solution.
We present a new method, self-ask, that further improves on chain of thought.
arXiv Detail & Related papers (2022-10-07T06:50:23Z) - Generative Context Pair Selection for Multi-hop Question Answering [60.74354009152721]
We propose a generative context selection model for multi-hop question answering.
Our proposed generative passage selection model has a better performance (4.9% higher than baseline) on adversarial held-out set.
arXiv Detail & Related papers (2021-04-18T07:00:48Z)
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.