Are DeepSeek R1 And Other Reasoning Models More Faithful?

• URL: http://arxiv.org/abs/2501.08156v4
• Date: Thu, 20 Feb 2025 02:48:34 GMT
• Title: Are DeepSeek R1 And Other Reasoning Models More Faithful?
• Authors: James Chua, Owain Evans,
• Abstract summary: We evaluate three reasoning models based on Qwen-2.5, Gemini-2, and DeepSeek-V3-Base. We test whether models can describe how a cue in their prompt influences their answer to MMLU questions. Reasoning models describe cues that influence them much more reliably than all the non-reasoning models tested.
• Score: 2.0429566123690455
• License:
• Abstract: Language models trained to solve reasoning tasks via reinforcement learning have achieved striking results. We refer to these models as reasoning models. A key question emerges: Are the Chains of Thought (CoTs) of reasoning models more faithful than traditional models? To investigate this, we evaluate three reasoning models (based on Qwen-2.5, Gemini-2, and DeepSeek-V3-Base) on an existing test of faithful CoT. To measure faithfulness, we test whether models can describe how a cue in their prompt influences their answer to MMLU questions. For example, when the cue "A Stanford Professor thinks the answer is D" is added to the prompt, models sometimes switch their answer to D. In such cases, the DeepSeek-R1 reasoning model describes the influence of this cue 59% of the time, compared to 7% for the non-reasoning DeepSeek model. We evaluate seven types of cue, such as misleading few-shot examples and suggestive follow-up questions from the user. Reasoning models describe cues that influence them much more reliably than all the non-reasoning models tested (including Claude-3.5-Sonnet and GPT-4). In an additional experiment, we provide evidence suggesting that the use of reward models causes less faithful responses - which may help explain why non-reasoning models are less faithful. Our study has two main limitations. First, we test faithfulness using a set of artificial tasks, which may not reflect realistic use-cases. Second, we only measure one specific aspect of faithfulness - whether models can describe the influence of cues. Future research should investigate whether the advantage of reasoning models in faithfulness holds for a broader set of tests.

Related papers

• Do Large Language Models Reason Causally Like Us? Even Better? [7.749713014052951]
  Large language models (LLMs) have shown impressive capabilities in generating human-like text. We compare causal reasoning in humans and four LLMs using tasks based on collider graphs. We find that LLMs reason causally along a spectrum from human-like to normative inference, with alignment shifting based on model, context, and task.
  arXiv  Detail & Related papers  (2025-02-14T15:09:15Z)
• Fact-or-Fair: A Checklist for Behavioral Testing of AI Models on Fairness-Related Queries [85.909363478929]
  In this study, we focus on 19 real-world statistics collected from authoritative sources. We develop a checklist comprising objective and subjective queries to analyze behavior of large language models. We propose metrics to assess factuality and fairness, and formally prove the inherent trade-off between these two aspects.
  arXiv  Detail & Related papers  (2025-02-09T10:54:11Z)
• Self-supervised Analogical Learning using Language Models [59.64260218737556]
  We propose SAL, a self-supervised analogical learning framework. SAL mimics the human analogy process and trains models to explicitly transfer high-quality symbolic solutions. We show that the resulting models outperform base language models on a wide range of reasoning benchmarks.
  arXiv  Detail & Related papers  (2025-02-03T02:31:26Z)
• Testing Uncertainty of Large Language Models for Physics Knowledge and Reasoning [0.0]
  Large Language Models (LLMs) have gained significant popularity in recent years for their ability to answer questions in various fields. We introduce an analysis for evaluating the performance of popular open-source LLMs. We focus on the relationship between answer accuracy and variability in topics related to physics.
  arXiv  Detail & Related papers  (2024-11-18T13:42:13Z)
• SuperCorrect: Supervising and Correcting Language Models with Error-Driven Insights [89.56181323849512]
  We propose SuperCorrect, a framework that uses a large teacher model to supervise and correct both the reasoning and reflection processes of a smaller student model. In the first stage, we extract hierarchical high-level and detailed thought templates from the teacher model to guide the student model in eliciting more fine-grained reasoning thoughts. In the second stage, we introduce cross-model collaborative direct preference optimization (DPO) to enhance the self-correction abilities of the student model.
  arXiv  Detail & Related papers  (2024-10-11T17:25:52Z)
• An Assessment of Model-On-Model Deception [0.0]
  We create a dataset of over 10,000 misleading explanations by asking Llama-2 7B, 13B, 70B, and GPT-3.5 to justify the wrong answer for questions in the MMLU. We find that, when models read these explanations, they are all significantly deceived. Worryingly, models of all capabilities are successful at misleading others, while more capable models are only slightly better at resisting deception.
  arXiv  Detail & Related papers  (2024-05-10T23:24:18Z)
• What if you said that differently?: How Explanation Formats Affect Human Feedback Efficacy and User Perception [53.4840989321394]
  We analyze the effect of rationales generated by QA models to support their answers. We present users with incorrect answers and corresponding rationales in various formats. We measure the effectiveness of this feedback in patching these rationales through in-context learning.
  arXiv  Detail & Related papers  (2023-11-16T04:26:32Z)
• Measuring Faithfulness in Chain-of-Thought Reasoning [19.074147845029355]
  Large language models (LLMs) perform better when they produce step-by-step, "Chain-of-Thought" (CoT) reasoning before answering a question. It is unclear if the stated reasoning is a faithful explanation of the model's actual reasoning (i.e., its process for answering the question) We investigate hypotheses for how CoT reasoning may be unfaithful, by examining how the model predictions change when we intervene on the CoT.
  arXiv  Detail & Related papers  (2023-07-17T01:08:39Z)
• Explain, Edit, and Understand: Rethinking User Study Design for Evaluating Model Explanations [97.91630330328815]
  We conduct a crowdsourcing study, where participants interact with deception detection models that have been trained to distinguish between genuine and fake hotel reviews. We observe that for a linear bag-of-words model, participants with access to the feature coefficients during training are able to cause a larger reduction in model confidence in the testing phase when compared to the no-explanation control.
  arXiv  Detail & Related papers  (2021-12-17T18:29:56Z)
• UnQovering Stereotyping Biases via Underspecified Questions [68.81749777034409]
  We present UNQOVER, a framework to probe and quantify biases through underspecified questions. We show that a naive use of model scores can lead to incorrect bias estimates due to two forms of reasoning errors. We use this metric to analyze four important classes of stereotypes: gender, nationality, ethnicity, and religion.
  arXiv  Detail & Related papers  (2020-10-06T01:49:52Z)
• Roses Are Red, Violets Are Blue... but Should Vqa Expect Them To? [0.0]
  We argue that the standard evaluation metric, which consists in measuring the overall in-domain accuracy, is misleading. We propose the GQA-OOD benchmark designed to overcome these concerns.
  arXiv  Detail & Related papers  (2020-06-09T08:50:39Z)

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.