Self-Contradictory Reasoning Evaluation and Detection
- URL: http://arxiv.org/abs/2311.09603v4
- Date: Mon, 21 Oct 2024 04:16:09 GMT
- Title: Self-Contradictory Reasoning Evaluation and Detection
- Authors: Ziyi Liu, Soumya Sanyal, Isabelle Lee, Yongkang Du, Rahul Gupta, Yang Liu, Jieyu Zhao,
- Abstract summary: We investigate self-contradictory (Self-Contra) reasoning, where the model reasoning does not support its answers.
We find that LLMs often contradict themselves in reasoning tasks involving contextual information understanding or commonsense.
We find that GPT-4 can detect Self-Contra with a 52.2% F1 score, much lower compared to 66.7% for humans.
- Score: 31.452161594896978
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: In a plethora of recent work, large language models (LLMs) demonstrated impressive reasoning ability, but many proposed downstream reasoning tasks only focus on final answers. Two fundamental questions persist: 1) how consistent is the reasoning, and 2) can models detect unreliable reasoning? In this paper, we investigate self-contradictory (Self-Contra) reasoning, where the model reasoning does not support its answers. To answer 1), we define and assess the Self-Contra rate across three datasets and delve into finer-grained categories of Self-Contra reasoning. We find that LLMs often contradict themselves in reasoning tasks involving contextual information understanding or commonsense. The model may generate correct answers by taking shortcuts in reasoning or overlooking contextual evidence, leading to compromised reasoning. For 2), we task the state-of-the-art model GPT-4 with identifying Self-Contra reasoning and finer-grained fallacies. We find that finer-grained categories enhanced detection can improve GPT-4's ability to detect Self-Contra. However, it is only able to detect Self-Contra with a 52.2% F1 score, much lower compared to 66.7% for humans. Our results indicate that current LLMs lack the robustness necessary for reliable reasoning and we emphasize the urgent need for establishing best practices in comprehensive reasoning evaluations beyond pure performance-based metrics.
Related papers
- Thinking About Thinking: SAGE-nano's Inverse Reasoning for Self-Aware Language Models [0.0]
Large Language Models (LLMs) have demonstrated remarkable capabilities at solving complex reasoning tasks with Chain-of-Thought prompting.<n>We introduce textbfinverse reasoning, a novel paradigm enabling LLMs to decompose and explain their own reasoning chains post-hoc.<n>Our work creates new avenues for transparent AI systems and closes significant gaps in AI safety, education, and scientific discovery.
arXiv Detail & Related papers (2025-06-30T09:53:41Z) - Lost at the Beginning of Reasoning [82.18834329384514]
We show that the first reasoning step exerts a disproportionately large influence on the final prediction.<n>We propose an efficient sampling strategy that leverages a reward model to identify and retain high-quality first reasoning steps.<n>We introduce a new benchmark specifically constructed with deliberately flawed first reasoning steps to systematically evaluate model self-correction capabilities.
arXiv Detail & Related papers (2025-06-27T09:53:57Z) - What's Missing in Vision-Language Models? Probing Their Struggles with Causal Order Reasoning [26.671128120554457]
causal reasoning is fundamental to solving complex high-level reasoning tasks.<n>Existing benchmarks often include a mixture of reasoning questions.<n>We introduce VQA-Causal and VCR-Causal to isolate and rigorously evaluate causal reasoning abilities.
arXiv Detail & Related papers (2025-06-01T07:17:46Z) - Revisiting Overthinking in Long Chain-of-Thought from the Perspective of Self-Doubt [74.35891434097053]
Reasoning Large Language Models (RLLMs) have demonstrated impressive performance on complex tasks.<n>They often exhibit overthinking -- performing unnecessary reasoning steps even after arriving at the correct answer.<n>We present a quantitative analysis of overthinking from the perspective of self-doubt.<n>We introduce a simple and effective prompting method to reduce the model's over-reliance on input questions.
arXiv Detail & Related papers (2025-05-29T14:30:02Z) - CoThink: Token-Efficient Reasoning via Instruct Models Guiding Reasoning Models [56.40065909544213]
Large language models (LLMs) benefit from increased test-time compute, a phenomenon known as test-time scaling.<n>However, reasoning-optimized models often overthink even simple problems, producing excessively verbose outputs and leading to low token efficiency.<n>We identify two key causes of this verbosity: (1) reinforcement learning reduces the information density of forward reasoning, and (2) backward chain-of thought training encourages redundant and often unnecessary verification steps.
arXiv Detail & Related papers (2025-05-28T06:24:45Z) - BARREL: Boundary-Aware Reasoning for Factual and Reliable LRMs [87.24843751412783]
We propose BARREL, a framework that promotes concise and boundary-aware factual reasoning.<n>Our experiments show that BARREL-training increases the reliability of DeepSeek-R1-Distill-Llama-8B from 39.33% to 61.48%.
arXiv Detail & Related papers (2025-05-18T07:27:34Z) - Beyond the Last Answer: Your Reasoning Trace Uncovers More than You Think [51.0691253204425]
We analyze intermediate reasoning steps, termed subthoughts, to answer two questions: Does the final answer reliably represent the model's optimal conclusion?
Our approach involves segmenting a reasoning trace into sequential subthoughts based on linguistic cues.
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.
arXiv Detail & Related papers (2025-04-29T12:39:07Z) - Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs [28.565225092457897]
Reinforcement learning can drive self-improvement in language models on verifiable tasks.
We find that Qwen-2.5-3B far exceeds Llama-3.2-3B under identical RL training for the game of Countdown.
Our study reveals that Qwen naturally exhibits these reasoning behaviors, whereas Llama initially lacks them.
arXiv Detail & Related papers (2025-03-03T08:46:22Z) - Information Re-Organization Improves Reasoning in Large Language Models [22.2946033364035]
We propose an information re-organization (InfoRE) method to enhance the reasoning ability of large language models (LLMs)
Our method involves extracting logical relationships from the contextual content, such as documents or paragraphs, and subsequently pruning redundant content to minimize noise.
To demonstrate the effectiveness of our approach in improving the reasoning ability, we conduct experiments using Llama2-70B, GPT-3.5, and GPT-4 on various contextually aware multi-hop reasoning tasks.
arXiv Detail & Related papers (2024-04-22T08:47:27Z) - Mitigating Misleading Chain-of-Thought Reasoning with Selective Filtering [59.495717939664246]
Large language models have manifested remarkable capabilities by leveraging chain-of-thought (CoT) reasoning techniques to solve intricate questions.
We propose a novel approach called the selective filtering reasoner (SelF-Reasoner) that assesses the entailment relationship between the question and the candidate reasoning chain.
SelF-Reasoner improves the fine-tuned T5 baseline consistently over the ScienceQA, ECQA, and LastLetter tasks.
arXiv Detail & Related papers (2024-03-28T06:28:35Z) - LogicAsker: Evaluating and Improving the Logical Reasoning Ability of Large Language Models [63.14196038655506]
We introduce LogicAsker, a novel approach for evaluating and enhancing the logical reasoning capabilities of large language models (LLMs)
Our methodology reveals significant gaps in LLMs' learning of logical rules, with identified reasoning failures ranging from 29% to 90% across different models.
We leverage these findings to construct targeted demonstration examples and fine-tune data, notably enhancing logical reasoning in models like GPT-4o by up to 5%.
arXiv Detail & Related papers (2024-01-01T13:53:53Z) - A Closer Look at the Self-Verification Abilities of Large Language Models in Logical Reasoning [73.77088902676306]
We take a closer look at the self-verification abilities of large language models (LLMs) in the context of logical reasoning.
Our main findings suggest that existing LLMs could struggle to identify fallacious reasoning steps accurately and may fall short of guaranteeing the validity of self-verification methods.
arXiv Detail & Related papers (2023-11-14T07:13:10Z) - How susceptible are LLMs to Logical Fallacies? [5.723715910568911]
We present LOGICOM, a diagnostic benchmark to assess the robustness of Large Language Models against logical fallacies.
We use this benchmark to evaluate the performance of GPT-3.5 and GPT-4 using a dataset containing controversial topics.
Our findings indicate that both GPT-3.5 and GPT-4 can adjust their opinion through reasoning.
arXiv Detail & Related papers (2023-08-18T23:07:29Z) - Question Decomposition Improves the Faithfulness of Model-Generated
Reasoning [23.34325378824462]
Large language models (LLMs) are difficult to verify the correctness and safety of their behavior.
One approach is to prompt LLMs to externalize their reasoning, by having them generate step-by-step reasoning as they answer a question.
This approach relies on the stated reasoning faithfully reflecting the model's actual reasoning, which is not always the case.
Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT.
arXiv Detail & Related papers (2023-07-17T00:54:10Z) - Language Models with Rationality [57.37201135072838]
Large language models (LLMs) are proficient at question-answering (QA)
It is not always clear how (or even if) an answer follows from their latent "beliefs"
arXiv Detail & Related papers (2023-05-23T17:04:25Z) - Consistency Analysis of ChatGPT [65.268245109828]
This paper investigates the trustworthiness of ChatGPT and GPT-4 regarding logically consistent behaviour.
Our findings suggest that while both models appear to show an enhanced language understanding and reasoning ability, they still frequently fall short of generating logically consistent predictions.
arXiv Detail & Related papers (2023-03-11T01:19:01Z) - Faithful Reasoning Using Large Language Models [12.132449274592668]
We show how LMs can be made to perform faithful multi-step reasoning via a process whose causal structure mirrors the underlying logical structure of the problem.
Our approach works by chaining together reasoning steps, where each step results from calls to two fine-tuned LMs.
We demonstrate the effectiveness of our model on multi-step logical deduction and scientific question-answering, showing that it outperforms baselines on final answer accuracy.
arXiv Detail & Related papers (2022-08-30T13:44:41Z) - Evaluate Confidence Instead of Perplexity for Zero-shot Commonsense
Reasoning [85.1541170468617]
This paper reconsiders the nature of commonsense reasoning and proposes a novel commonsense reasoning metric, Non-Replacement Confidence (NRC)
Our proposed novel method boosts zero-shot performance on two commonsense reasoning benchmark datasets and further seven commonsense question-answering datasets.
arXiv Detail & Related papers (2022-08-23T14:42:14Z)
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.