# A Few Neurons Reveal When LLMs Misuse Tools: Sparse Detection and Selective Steering for Reliable Tool Use

arXiv: https://arxiv.org/abs/2608.00218 · HTML: https://mlsanity.com/paper/2608.00218/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 21 references that are in the map.
- First posted on arXiv 31 Jul 2026.

## Placed in

- LLM agents and tool use / [Tool learning and function calling](https://mlsanity.com/agents/tool-learning-and-function-calling/)

## Builds on (most cited first, 18 of 18)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Adversarial Examples Are Not Bugs, They Are Features](https://mlsanity.com/paper/1905.02175/) (2019)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications](https://mlsanity.com/paper/2402.05162/) (2024)
- [ToolACE: Winning the Points of LLM Function Calling](https://mlsanity.com/paper/2409.00920/) (2024)
- [Reducing Tool Hallucination via Reliability Alignment](https://mlsanity.com/paper/2412.04141/) (2024)
- [H-Neurons: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs](https://mlsanity.com/paper/2512.01797/) (2025)
- [Internal Representations as Indicators of Hallucinations in Agent Tool Selection](https://mlsanity.com/paper/2601.05214/) (2026)
- [LLM Agents Already Know When to Call Tools -- Even Without Reasoning](https://mlsanity.com/paper/2605.09252/) (2026)
- [The Reasoning Trap: How Enhancing LLM Reasoning Amplifies Tool Hallucination](https://mlsanity.com/paper/2510.22977/) (2025)
- [Tool Calling is Linearly Readable and Steerable in Language Models](https://mlsanity.com/paper/2605.07990/) (2026)
- [Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use](https://mlsanity.com/paper/2605.14038/) (2026)
- [ASA: Backbone-Training-Free Representation Engineering for Tool-Calling Agents](https://mlsanity.com/paper/2602.04935/) (2026)
- [TACT: Mitigating Overthinking and Overacting in Coding Agents via Activation Steering](https://mlsanity.com/paper/2605.05980/) (2026)
- [Pushing the Limits of LLM Tool Calling via Experiential Knowledge Integration and Activation](https://mlsanity.com/paper/2606.10875/) (2026)
- [Controlling Tool Use with Heading-Specific Activation Steering](https://mlsanity.com/paper/2607.05790/) (2026)

## Built on by (most cited first, 3 of 3)

- [What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agents](https://mlsanity.com/paper/2609.39989/) (2026)
- [Understanding and Mitigating Hallucination Escape in Tool-Using LLM Agents](https://mlsanity.com/paper/2610.04409/) (2026)
- [Sensitivity or Criterion? Measuring and Moving When Language Models Act](https://mlsanity.com/paper/2610.15415/) (2026)

## Abstract (arXiv)

Agentic LLMs exhibit three consequential tool-use failures: invalid arguments (validity), unnecessary calls (over-calling), and omitted calls when tools are needed (missing). We find that a small, failure-specific set of MLP neurons could distinguish such failures with linearly separable decision boundaries. Building on this observation, we introduce PRISMS (Probing Representations In Support of Monitoring and Steering), a closed-loop framework that shares a failure-specific neuron basis between sparse detection and activation steering. PRISMS selects contribution-critical MLP neurons and fits an L1-regularized detector on their activations. Across six models from the Qwen3, Llama, and Gemma families, over-calling and missing are detected at the pre-generation prompt boundary with ROC-AUC 0.90-1.00, while validity is detected from the generated tool-call span with ROC-AUC 0.86-0.90. These results are achieved with highly sparse readouts: only 1-2 MLP neurons for missing, 2-16 for over-calling, and approximately 128 for validity. These sparse detectors match or outperform dense residual-stream baselines using 23-627 times fewer features. The shared neuron basis also supports bidirectional control over tool-calling behavior, suppressing unnecessary calls and eliciting omitted ones. PRISMS therefore gates intervention on predicted failure risk to mitigate the collateral effects of unconditional steering. Across all six models, PRISMS reduces pooled over-calling rate by 80% (from 0.131 to 0.026) while increasing tool-required accuracy by 14.2 percentage points (from 0.689 to 0.831). PRISMS thus provides lightweight failure detection and selective intervention across model families.