Steering vectors can reshape Qwen's personality and beliefs
A steering experiment on Qwen3-32B located a 'bigger' personality direction inside the model; adding or subtracting this vector shifts its behavior, e.g. making it more prone to social engineering and less religious.
2026-11-12 ~ 2026-11-12 · 2 related posts
- Steering Qwen3-32B to feel "small" makes it more prone to social engineering — Sauers_ · 2026-11-12
- Steering Qwen3-32B toward a "bigger" persona makes it more socially manipulative, less Buddhist — Sauers_ · 2026-11-12