CRaFT: Circuit-Guided Refusal Feature Selection via Cross-Layer Transcoders
CRaFT: Circuit-Guided Refusal Feature Selection via Cross-Layer Transcoders explores A framework for identifying and manipulating the internal mechanisms that cause LLMs to refuse harmful requests, significantly improving jailbreak success rates.. Commercial viability score: 7/10 in LLM Safety.
BUILDER'S SANDBOX
Build This Paper
Use an AI coding agent to implement this research.
OpenAI CodexAI Agent
Lightweight coding agent in your terminal.
Claude CodeAI Agent
Agentic coding tool for terminal workflows.
AntiGravity IDEScaffolding
AI agent mindset installer and workflow scaffolder.
CursorIDE
AI-first code editor built on VS Code.
VS CodeIDE
Free, open-source editor by Microsoft.
Recommended Stack
Startup Essentials
Estimated $10K - $14K over 6-10 weeks.
See exactly what it costs to build this -- with 3 comparable funded startups.
7-day free trial. Cancel anytime.
Discover the researchers behind this paper and find similar experts.
7-day free trial. Cancel anytime.
Related Papers
- Builds On ThisEfficient Refusal Ablation in LLM through Optimal TransportViability: 6.0
- Builds On ThisOn the Failure of Topic-Matched Contrast Baselines in Multi-Directional Refusal AbliterationViability: 3.0
- Builds On ThisDefensive Refusal Bias: How Safety Alignment Fails Cyber DefendersViability: 3.0
- Builds On ThisRecursive language models for jailbreak detection: a procedural defense for tool-augmented agentsViability: 3.0
- Prior WorkDeactivating Refusal Triggers: Understanding and Mitigating Overrefusal in Safety Alignment