mlsanityData to 7 Oct 2026

INTENT-AS-A-TOOL Makes it Easy to Track Agentic Misalignment

arXiv 2608.27348 ↗First posted 27 Aug 2026cs.CLcited by 0

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

As large language models (LLMs) are deployed as autonomous agents, safety failures increasingly involve consequential actions. We study agentic misalignment, where agents take harmful actions under goal conflicts and pressures. Using chain-of-thought (CoT) monitoring, we find that harmful execution is often preceded by intent signals in reasoning. However, post-hoc CoT labels are too coarse to show how intent changes during generation. We introduce INTENT-AS-A-TOOL, an approach that adds intent-targeted tools to give the model a dedicated channel for expressing commitment to a target behavior. The probability of calling an intent tool provides a judge-free, fine-grained signal of the model's tendency to pursue that behavior. Our results show that INTENT-AS-A-TOOL complements CoT monitoring, expands post-hoc CoT labels into dense trajectories, and identifies critical steps for online intervention. These findings suggest that action preferences are useful for tracking agentic misalignment during reasoning. Our code and data are accessible: https://github.com/RebeccaZhang22/intent-as-a-tool.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 33 of its 49 references that are in the map. First posted on arXiv 27 Aug 2026.

Builds on

33 of 49 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025Agentic Misalignment: How LLMs Could Be Insider Threats5 Oct
  2. 2025Reliable Weak-to-Strong Monitoring of LLM Agents26 Aug
  3. 2024AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents11 Oct
  4. 2023Identifying the Risks of LM Agents with an LM-Emulated Sandbox25 Sep

Other references · 29

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  2. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  3. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
  4. 2023AgentBench: Evaluating LLMs as Agentscited by 1,211
  5. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 965
  6. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 960
  7. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deepcited by 450
  8. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscationcited by 338
  9. 2024Improving Alignment and Robustness with Circuit Breakerscited by 302
  10. 2024Frontier Models are Capable of In-context Schemingcited by 252
  11. 2023Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theorycited by 218
  12. 2024Chain-of-Thought Reasoning Without Promptingcited by 208
17 more
  1. 2023AI Control: Improving Safety Despite Intentional Subversioncited by 207
  2. 2024PrivacyLens: Evaluating Privacy Norm Awareness of Language Models in Actioncited by 161
  3. 2025Thought Anchors: Which LLM Reasoning Steps Matter?cited by 126
  4. 2025Large Language Models Often Know When They Are Being Evaluatedcited by 105
  5. 2024Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMscited by 78
  6. 2025Understanding Reasoning in Thinking Language Models via Steering Vectorscited by 69
  7. 2025Effectively Controlling Reasoning Models through Thinking Interventioncited by 63
  8. 2025CoT Red-Handed: Stress Testing Chain-of-Thought Monitoringcited by 49
  9. 2024Monitoring Latent World States in Language Models with Propositional Probescited by 39
  10. 2025Steering Evaluation-Aware Language Models to Act Like They Are Deployedcited by 30
  11. 2025The Hawthorne Effect in Reasoning Models: Evaluating and Steering Test Awarenesscited by 29
  12. 2025Is It Thinking or Cheating? Detecting Implicit Reward Hacking by Measuring Reasoning Effortcited by 25
  13. 2025SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignmentcited by 24
  14. 2026Reasoning Models Struggle to Control their Chains of Thoughtcited by 22
  15. 2025How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?cited by 11
  16. 2025Towards Safe Reasoning in Large Reasoning Models via Corrective Interventioncited by 11
  17. 2026The Model's Tell: Measuring Context-Leakage Attack Signals with Behavior Gaugescited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026When Users Change Their Minds: Measuring and Repairing Intent Drift in LLM Agents
  2. 2026Learning When to Act or Refuse: Guarding Agentic Reasoning Models for Safe Multi-Step Tool Use
  3. 2025AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models
  4. 2026When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents
  5. 2026AgentBoundary: Counterfactual Evaluation of Safety in Tool-Using LLM Agents
  6. 2026Model Forensics: Investigating Whether Concerning Behavior Reflects Misalignment
  7. 2026Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation
  8. 2026PASTABench: Proactive Assessment of Sequential Trajectories for Agent Safety