Do as I Say, Not as I Do: Instruction-Induction Conflict in LLMs

• URL: http://arxiv.org/abs/2605.20382v1
• Date: Tue, 19 May 2026 18:32:20 GMT
• Title: Do as I Say, Not as I Do: Instruction-Induction Conflict in LLMs
• Abstract summary: Language models are trained to follow instructions, but they are also powerful pattern completers.<n>We construct conversations in which a user instruction to behave in a target way T is opposed by N hardcoded assistant turns demonstrating a competing pattern P.<n>We measure instruction-following (IF) rates in this setting, across 13 models and 16 different instructions, for up to 50 turns.
• Score: 0.17188280334580197
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Language models are trained to follow instructions, but they are also powerful pattern completers. What happens when these two objectives conflict? We construct conversations in which a user instruction to behave in a target way T (e.g., always output a specific token, answer in a particular language, or adopt a persona) is opposed by N hardcoded assistant turns demonstrating a competing pattern P. We then measure instruction-following (IF) rates in this setting, across 13 models and 16 different instructions, for up to 50 turns. Average instruction-following rates range from 1% to 99% across models, largely uncorrelated with standard capability benchmarks. The transition from instruction-following to pattern-following is universal but highly model-dependent. Robustness is modulated both by instruction content, with models resisting induction longer when instructions align with their trained value priors, and by output format, with diverse multi-token responses proving substantially more resistant than single-token outputs. Chain-of-thought reasoning improves robustness but does not eliminate susceptibility, and can produce dissociation between correct deliberation and incorrect output. When asked to predict their behavior in this setting, models achieve 83.5% accuracy on average but systematically underestimate their own resistance to induction pressure. These results suggest that instruction-following remains brittle under induction pressure even for otherwise capable models, and that output diversity, rather than semantic engagement with the input, is the primary factor predicting robustness.

Related papers

• When Tokenization is Secretly Output Supervision [10.546694596577057]
  Tokenization in language models is treated by default as an input preprocessing decision.<n>In autoregressive models, tokenizer granularity determines what the model must resolve in a single forward pass.<n>This affects both the difficulty of the learning problem and the representations that emerge inside the model.
  arXiv  Detail & Related papers  (2026-09-01T15:16:19Z)
• Task Competence Is Not Instruction Following: Evaluating Instruction-Conflicting Behavior in Small Language Models [3.6463026920885824]
  We evaluate instruction-tuned Qwen models across sizes.<n>Both standard accuracy and instruction following generally improve with scale.<n>Small models stay competent yet routinely ignore the non-standard instruction, while larger models show a clear gap between the two settings.
  arXiv  Detail & Related papers  (2026-07-21T22:24:18Z)
• Right or Wrong, Models Comply: Directional Blindness in LLM Moral Judgment [13.021647720314016]
  We introduce Compliance Asymmetry, a bidirectional diagnostic that compares beneficial output change under helpful nudges with harmful nudges.<n>We find that this selectivity differs in factual and moral judgments.<n>Chain-of-thought prompting amplifies helpful and harmful compliance together, while identity-based prompting suppresses both by nearly identical margins.
  arXiv  Detail & Related papers  (2026-06-12T02:23:38Z)
• 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)
• Prototype-Based Dynamic Steering for Large Language Models [3.90727941420584]
  Prototype-Based Dynamic Steering (PDS) is a test-time method that amplifies large language model (LLM) reasoning without adding or altering instructions.<n>We introduce "reasoning prototypes" by clustering activation differences between Chain-of-Thought (CoT) and neutral prompts.<n>PDS consistently improves accuracy without fine-tuning or prompt engineering.
  arXiv  Detail & Related papers  (2025-10-07T01:34:28Z)
• Sequential Diffusion Language Models [110.06562906987052]
  Diffusion language models (DLMs) have strong theoretical efficiency but are limited by fixed-length decoding and incompatibility with key-value caches.<n>We introduce Next Sequence Prediction (NSP), which unifies next-token and next-block prediction.<n>We propose Sequential Diffusion Language Model (SDLM), which can retrofit pre-trained autoregressive language models (ALMs) at minimal cost.
  arXiv  Detail & Related papers  (2025-09-28T17:59:15Z)
• Adversarial Manipulation of Reasoning Models using Internal Representations [1.308812559547533]
  Reasoning models generate chain-of-thought (CoT) tokens before their final output.<n>We identify a linear direction in activation space during CoT token generation that predicts whether the model will refuse or comply.<n>We show that intervening only on CoT token activations suffices to control final outputs, and that incorporating this direction into prompt-based attacks improves success rates.
  arXiv  Detail & Related papers  (2025-07-03T20:51:32Z)
• Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations [83.93566096400723]
  We find that instruction-tuned models retain up to 93.4% of their original performance when given a randomly sampled tokenization.<n>Character-level segmentation improves string manipulation and code understanding tasks by up to +14%.<n>Right-aligned digit grouping enhances large-number arithmetic by +33%.
  arXiv  Detail & Related papers  (2025-06-23T18:02:26Z)
• The Price of a Second Thought: On the Evaluation of Reasoning Efficiency in Large Language Models [54.88805865447848]
  We show that instruct models achieve higher efficiency overall, and problem difficulty affects efficiency.<n>We propose COTHINK, a simple two-stage pipeline: an instruct model drafts a brief outline, and a thinking model expands it.<n>On GSM8K, MATH500, and AIME24, COTHINK cuts token usage by 21.1% while keeping accuracy on four thinking models, and remains competitive with strong efficiency baselines.
  arXiv  Detail & Related papers  (2025-05-28T06:24:45Z)
• Language Models Can Predict Their Own Behavior [29.566208688211876]
  Language models (LMs) can exhibit specific behaviors,' such as a failure to follow alignment training, that we hope to detect and react to during deployment.<n>We show that probes trained on the internal representation of input tokens alone can predict a wide range of eventual behaviors over the entire output sequence.<n>An early warning system built on the probes reduces jailbreaking by 91%.
  arXiv  Detail & Related papers  (2025-02-18T23:13:16Z)
• Instruction Position Matters in Sequence Generation with Large Language Models [67.87516654892343]
  Large language models (LLMs) are capable of performing conditional sequence generation tasks, such as translation or summarization. We propose enhancing the instruction-following capability of LLMs by shifting the position of task instructions after the input sentences.
  arXiv  Detail & Related papers  (2023-08-23T12:36:57Z)
• Enhancing Large Language Models Against Inductive Instructions with Dual-critique Prompting [55.15697111170836]
  This paper reveals the behaviors of large language models (LLMs) towards textitinductive instructions and enhance their truthfulness and helpfulness accordingly. After extensive human and automatic evaluations, we uncovered a universal vulnerability among LLMs in processing inductive instructions. We identify that different inductive styles affect the models' ability to identify the same underlying errors, and the complexity of the underlying assumptions also influences the model's performance.
  arXiv  Detail & Related papers  (2023-05-23T06:38:20Z)
• Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition [62.83832841523525]
  We propose a fast and accurate parallel transformer, termed Paraformer. It accurately predicts the number of output tokens and extract hidden variables. It can attain comparable performance to the state-of-the-art AR transformer, with more than 10x speedup.
  arXiv  Detail & Related papers  (2022-06-16T17:24:14Z)
• AvgOut: A Simple Output-Probability Measure to Eliminate Dull Responses [97.50616524350123]
  We build dialogue models that are dynamically aware of what utterances or tokens are dull without any feature-engineering. The first model, MinAvgOut, directly maximizes the diversity score through the output distributions of each batch. The second model, Label Fine-Tuning (LFT), prepends to the source sequence a label continuously scaled by the diversity score to control the diversity level. The third model, RL, adopts Reinforcement Learning and treats the diversity score as a reward signal.
  arXiv  Detail & Related papers  (2020-01-15T18:32:06Z)

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.