Layer 26 · Neuron 279 · boosted

L26N279+1

Word quirks:

“<”“ subscript”“$_”“.sub”“�”“sub”“ivio”“_sub”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

The reply didn't hold together

Isub<

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L26N4523-1🕒	0.83	“ суб”“子”“sub”
2. L27N6800-1🕒	0.83	“ sub”“sub”“ Sub”
3. L28N3725+1🕒	0.83	“ sub”“ Sub”“sub”
4. L27N6433-1💤	0.83	“ sub”“ Sub”“ SV”
5. L29N4614-1🕒	0.82	“ Sub”“sub”“ sub”
6. L27N5372+1💤	0.82	“子”“ Sub”“ sub”
7. L28N4986+1🕒	0.82	“sub”“ Sub”“ суб”
8. L29N6553+1🕒	0.82	“ sub”“sub”“ Sub”
9. L26N9040-1🕒	0.82	“ sub”“sub”“ Sub”
10. L25N3485-1🕒	0.81	“Sub”“ Sub”“-sub”

Shadow specimen, apart from the ten: L26N279-1💤, 0.05.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 26, neuron 279, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L26N279+1 t0-v0.1	1.222	0.64	1.270	2.06	Top boosted tokens ↑
L26N279-1💤 shadow specimen t0-v0.1	0.625	0.73	0.620	1.21	“如因”“超高”“/hash”“ливо”“在高”“-inner”“当面”“ attenu”

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.