# CRaFT: Circuit-Guided Refusal Feature Selection via Cross-Layer Transcoders

arXiv: https://arxiv.org/abs/2604.01604 · HTML: https://mlsanity.com/paper/2604.01604/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 10 of its 15 references that are in the map.
- First posted on arXiv 2 Apr 2026.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)
- Mechanistic interpretability / [Steering and probing for LLM safety](https://mlsanity.com/mechinterp/steering-and-probing-for-llm-safety/)

## Builds on (most cited first, 10 of 10)

- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Red Teaming Language Models with Language Models](https://mlsanity.com/paper/2202.03286/) (2022)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs](https://mlsanity.com/paper/2401.06373/) (2024)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [A Wolf in Sheep's Clothing: Generalized Nested Jailbreak Prompts can Fool Large Language Models Easily](https://mlsanity.com/paper/2311.08268/) (2023)
- [Transcoders Find Interpretable LLM Feature Circuits](https://mlsanity.com/paper/2406.11944/) (2024)
- [Understanding Refusal in Language Models with Sparse Autoencoders](https://mlsanity.com/paper/2505.23556/) (2025)
- [Holistic Automated Red Teaming for Large Language Models through Top-Down Test Case Generation and Multi-turn Interaction](https://mlsanity.com/paper/2409.16783/) (2024)

## Built on by (most cited first, 1 of 1)

- [Circuit-Diff: Factual Edit-based Intervention Method for Localizing Knowledge in Attribution Graphs](https://mlsanity.com/paper/2609.23892/) (2026)

## Abstract (arXiv)

While modern LLMs are aligned to refuse harmful requests, it is essential to understand the underlying mechanistic basis of this refusal behavior for model safety analysis. For example, steering-based jailbreak attacks exploit this by identifying and manipulating sparse, neuron-like refusal features to bypass safety guardrails. Current feature selection methods primarily rely on how strongly features activate on harmful prompts. However, activation strength alone often captures superficial heuristics such as topic or lexical cues, rather than the true causal mechanisms. Thus, selecting refusal features requires measuring inter-feature relationships, rather than treating each feature as an isolated activation signal. Based on this insight, we propose CRaFT, a circuit-guided framework for identifying critical refusal features that directly govern the refusal decision. CRaFT leverages cross-layer transcoders to map the model's internal computations into a sparse feature circuit graph, where edges quantify inter-feature influences and their contributions to the final output logits. By aggregating the effects propagating along the paths to refusal, CRaFT effectively ranks the most influential features. Extensive evaluations across four jailbreak benchmarks show that CRaFT significantly improves average performance from 6.7% to 57.4% and generates more specific harmful completions compared to current SOTA methods.