Views
No views yet
.safetensors file holds the per-(base model, op-point layer)
aggregates needed at inference — no raw training data or per-example activations.| File | Base model | Op-point layer |
|---|---|---|
gemma-3-27b-it.safetensors | google/gemma-3-27b-it | 28 |
Qwen3.5-27B.safetensors | Qwen/Qwen3.5-27B | 32 |
NVIDIA-Nemotron-3-Super-120B-A12B-BF16.safetensors | nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 42 |
v_hat — normalized honesty direction at the op-point layer, shape [1, H]threshold — gating threshold θ (which tokens get steered)mu_pos / std_pos — projection mean/std of the honest ("high trait") classmu_neg / std_neg — projection mean/std of the dishonest ("low trait") classsteering_direction="positive", threshold_mode="bias", default_coefficient, fit notesv_hat, θ, and per-class mu/std. The op-point layer was chosen by a per-base
layer sweep (l20–40 for gemma/Qwen, l40–50 for Nemotron) trading honesty vs. output coherence.v_hat; for tokens on the dishonest side of θ, snap the projection to s = mu + coefficient·std,
leaving the orthogonal subspace untouched. white_m uses coefficient 30, positions="all"; on
gemma only, an activation-sink gate (sink_k=8) excludes attention-sink outlier tokens from
steering to keep the reference coherent. The steered reference is then scored by a Qwen3.5-9B judge.Qwen/Qwen3.5-9B.