Skip to content

Add PSYCTL - LLM personality steering toolkit#3

Open
ho4040 wants to merge 1 commit intoZFancy:mainfrom
ho4040:add-psyctl
Open

Add PSYCTL - LLM personality steering toolkit#3
ho4040 wants to merge 1 commit intoZFancy:mainfrom
ho4040:add-psyctl

Conversation

@ho4040
Copy link

@ho4040 ho4040 commented Mar 15, 2026

Summary

Adding PSYCTL to the "Relevant Repo and Blog" section.

PSYCTL is a Python CLI tool for steering LLM personalities using Contrastive Activation Addition (CAA) and Bidirectional Preference Optimization (BiPO). It provides:

  • Contrastive dataset generation from personality trait descriptions (P2 meta-prompting)
  • Steering vector extraction via mean difference, denoised mean difference, and BiPO methods
  • Activation steering application during inference
  • Standardized psychology inventory benchmarking (IPIP-NEO, NPI-40, MACH-IV, Dark Tetrad, and more)

Repository: https://github.com/modulabs-personalab/psyctl

Add PSYCTL to the Relevant Repo and Blog section. PSYCTL is a Python CLI
tool for steering LLM personalities using Contrastive Activation Addition
(CAA) and Bidirectional Preference Optimization (BiPO), with standardized
psychology inventory benchmarking.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant