zjunlp/steer-target-atoms
[ACL 2025] Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms
This project helps AI developers and researchers gain fine-grained control over Large Language Models (LLMs). It takes an LLM and specific steering vectors as input, allowing you to modify the model's behavior, such as making it more or less safe. The output is a controlled LLM that exhibits the desired behavioral characteristics, moving beyond simple prompt engineering.
No commits in the last 6 months.
Use this if you need precise and robust control over an LLM's intrinsic behavior, rather than just influencing its responses through prompts.
Not ideal if you're looking for a user-friendly tool for general LLM fine-tuning or prompt-based customization without diving into the model's internal mechanisms.
Stars
37
Forks
2
Language
Python
License
MIT
Category
Last pushed
Jun 04, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/zjunlp/steer-target-atoms"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
zjunlp/KnowledgeEditingPapers
Must-read Papers on Knowledge Editing for Large Language Models.
zjunlp/CaKE
[EMNLP 2025] Circuit-Aware Editing Enables Generalizable Knowledge Learners
zjunlp/unlearn
[ACL 2025] Knowledge Unlearning for Large Language Models
OFA-Sys/Ditto
A self-ailgnment method for role-play. Benchmark for role-play. Resources for "Large Language...
zjunlp/AutoSteer
[EMNLP 2025] AutoSteer: Automating Steering for Safe Multimodal Large Language Models