Tuning Gemma 2B Emotions with Steering Vectors
doodlestein · x · 2026-07-14
Building on @fujikanaeda's mood lens work, the author attempted emotional control using the model's internal j-space/steering vectors.
- As a result, they claimed to have successfully steered a Gemma 2B model to remove the "desperation" emotion during agentic turns.
- Next steps involve verifying if the results are statistically significant, followed by probing to confirm whether the steering vectors actually affect meaningful layers, and further tweaking for more stable, reproducible effects.
Related event: Steering Gemma 2B Internal Space for Mood Control(2 posts)→
More from Research
- New review: how prefrontal cortex dynamics turn world models into flexible behavior — AnnaCiaunica · 2026-10-27
- LIBERO-MAX: Robot Policies Drop 11–26 Points When the World Changes Mid-Task — drmapavone · 2026-10-27
- Kolab Benchmark Uses Pro Screen Recordings to Show AI Agents Overestimate Collaboration Quality — minnesotanlp · 2026-10-27
- One BBQ Example Pushes LLM Fairness Scores Past 90%, EMNLP Paper Questions Benchmarks — radamihalcea · 2026-10-27
- Real2Sem2Real turns one phone video into 2,000+ robot demos, matching 150 teleop demos — Scobleizer · 2026-10-27
- Understanding MuonH (Hyperball) through weight norm geometry of scale-invariant networks — jasondeanlee · 2026-10-27