Implicit Reasoning Steering via Concept Chaining

• URL: http://arxiv.org/abs/2607.14242v1
• Date: Wed, 15 Jul 2026 18:03:08 GMT
• Title: Implicit Reasoning Steering via Concept Chaining
• Abstract summary: We use natural-language text to bias a model toward a designated answer without explicit instructions, triggers, or direct answer cues.<n>We show that indirect, natural-looking text can systematically steer model predictions while remaining substantially less inferable than direct paraphrases.
• Score: 29.121206944212556
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models often appear to reason reliably, yet on many questions repeated sampling yields both correct and incorrect answers, revealing an underlying fragility in how final decisions are formed. We study whether this fragility can be exploited through implicit reasoning steering: using natural-language text to bias a model toward a designated answer without explicit instructions, triggers, or direct answer cues. Our approach, Concept Chaining, generates a short connection paragraph that links question entities to a target option through one or two intermediate concepts. We then continue pretraining a victim model on these connection paragraphs and evaluate whether its answer preference shifts on the original multiple-choice questions. Our results show that indirect, natural-looking text can systematically steer model predictions while remaining substantially less inferable than direct paraphrases, which shows that reasoning brittleness is not merely an evaluation artifact: it creates a practical channel through which latent biases can be amplified by ordinary-looking text to covertly redirect model decisions.

Related papers

• Clarify, Abstain or Answer? Strategising in Conversation with Belief-Augmented Generation [38.81368744366676]
  Large language models (LLMs) define a distribution over text, which can be viewed as a probabilistic representation of uncertainty.<n>We propose Belief-Augmented Generation (BAG): grounding LLMs in their own belief state via the prompt and letting them reason over these K samples to decide on a conversational strategy.<n>BAG improves QA accuracy across six models and yields strategy decisions more faithful to the belief state than prompt-only baselines.
  arXiv  Detail & Related papers  (2026-05-25T13:29:02Z)
• Decoding Answers Before Chain-of-Thought: Evidence from Pre-CoT Probes and Activation Steering [5.427346259545067]
  Chain-of-thought (CoT) has become central to scaling reasoning capabilities in large language models.<n>We show that instruction-tuned models often determine their answer before generating CoT.
  arXiv  Detail & Related papers  (2026-03-02T04:33:55Z)
• Analyzing Reasoning Consistency in Large Multimodal Models under Cross-Modal Conflicts [74.47786985522762]
  We identify a critical failure mode termed textual inertia, where models tend to blindly adhere to the erroneous text while neglecting conflicting visual evidence.<n>We propose the LogicGraph Perturbation Protocol that structurally injects perturbations into the reasoning chains of diverse LMMs.<n>Results reveal that models successfully self-correct in less than 10% of cases and predominantly succumb to blind textual error propagation.
  arXiv  Detail & Related papers  (2026-01-07T16:39:34Z)
• More Bias, Less Bias: BiasPrompting for Enhanced Multiple-Choice Question Answering [53.09478307383865]
  We introduce BiasPrompting, a novel inference framework for large language models (LLMs)<n>It guides LLMs to generate and critically evaluate reasoning across all plausible answer options before reaching a final prediction.<n>It demonstrates significant improvements in five widely used multiple-choice question answering benchmarks.
  arXiv  Detail & Related papers  (2025-11-25T09:01:08Z)
• From Disagreement to Understanding: The Case for Ambiguity Detection in NLI [3.5164356851161984]
  This position paper argues that annotation disagreement in Natural Language Inference (NLI) is not mere noise but often reflects meaningful variation.<n>We call for a shift toward ambiguity-aware NLI that first identifies ambiguous input pairs, classifies their types, and only then proceeds to inference.
  arXiv  Detail & Related papers  (2025-07-20T20:27:35Z)
• A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Models [58.32070787537946]
  Chain-of-thought (CoT) reasoning enhances performance of large language models.<n>We present the first comprehensive study of CoT faithfulness in large vision-language models.
  arXiv  Detail & Related papers  (2025-05-29T18:55:05Z)
• Threading the Needle: Reweaving Chain-of-Thought Reasoning to Explain Human Label Variation [60.18907916989796]
  Large Language Models (LLMs) generate chains of thought (CoTs) before giving the final answer.<n>We propose a novel pipeline enriched with linguistically-grounded discourse segmenters to extract supporting and opposing statements for each answer option.<n>We also propose a rank-based HLV evaluation framework that prioritizes the ranking of answers over exact scores.
  arXiv  Detail & Related papers  (2025-05-29T11:47:18Z)
• Reasoning-Grounded Natural Language Explanations for Language Models [2.7855886538423182]
  We propose a large language model explainability technique for obtaining faithful natural language explanations.<n>When converted to a sequence of tokens, the outputs of the reasoning process can become part of the model context.<n>We show that the proposed use of reasoning can also improve the quality of the answers.
  arXiv  Detail & Related papers  (2025-03-14T10:00:03Z)
• Abductive Commonsense Reasoning Exploiting Mutually Exclusive Explanations [118.0818807474809]
  Abductive reasoning aims to find plausible explanations for an event. Existing approaches for abductive reasoning in natural language processing often rely on manually generated annotations for supervision. This work proposes an approach for abductive commonsense reasoning that exploits the fact that only a subset of explanations is correct for a given context.
  arXiv  Detail & Related papers  (2023-05-24T01:35:10Z)
• Logical Satisfiability of Counterfactuals for Faithful Explanations in NLI [60.142926537264714]
  We introduce the methodology of Faithfulness-through-Counterfactuals. It generates a counterfactual hypothesis based on the logical predicates expressed in the explanation. It then evaluates if the model's prediction on the counterfactual is consistent with that expressed logic.
  arXiv  Detail & Related papers  (2022-05-25T03:40:59Z)
• A Semantic-based Method for Unsupervised Commonsense Question Answering [40.18557352036813]
  Unsupervised commonsense question answering is appealing since it does not rely on any labeled task data. We present a novel SEmantic-based Question Answering method (SEQA) for unsupervised commonsense question answering.
  arXiv  Detail & Related papers  (2021-05-31T08:21:52Z)
• Contrastive Explanations for Model Interpretability [77.92370750072831]
  We propose a methodology to produce contrastive explanations for classification models. Our method is based on projecting model representation to a latent space. Our findings shed light on the ability of label-contrastive explanations to provide a more accurate and finer-grained interpretability of a model's decision.
  arXiv  Detail & Related papers  (2021-03-02T00:36:45Z)

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.