Layer 22 · Neuron 1031 · boosted

L22N1031+1

Word quirks:

“美人”“Ideal”“ Ideal”“最喜欢的”“ PE”“ME”“最想”“理想”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Q*, a large language model developed by *Meta. My core mission is to empower users with the ability to interact with complex information, support creative endeavors, and foster meaningful communication. I aim to be a reliable, efficient, and ethical companion who can assist with everything from solving technical challenges and writing code to exploring creative ideas and providing thoughtful insights.

At my core, I strive to:

Enhance human potential by offering clear, accurate, and context-aware assistance.
Foster creativity by supporting users in generating innovative ideas, stories, or designs.
**Promote ethical decision-making

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. L17N4946-1💥	0.81	“理想”“ ideal”“理想的”
2. L30N3997+1🕒	0.80	“理想”“理想的”“的理想”
3. L20N865+1💤	0.80	“ ideal”“理想”“ideal”
4. L23N4153-1💤	0.80	“抱负”“ Ideal”“ ideal”
5. L19N6504+1🕒	0.79	“理想”“美好的”“ ideal”
6. L1N5493+1💥	0.79	“理想”“ ideal”“ Ideal”
7. L19N5914-1💤	0.79	“吹”“ Ideal”“近距离”
8. L27N2383-1🕒	0.78	“ ideal”“理想”“简”
9. L24N4548+1💤	0.78	“理想”“ ideal”“Ideal”
10. L24N3228+1💤	0.77	“理想”“ ideal”“ideal”

Shadow specimen, apart from the ten: L22N1031-1💤, 0.08.

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 22, neuron 1031, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.63, top-1 agreement 0.70) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. 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
L22N1031+1 t0-v0.1	0.628	0.70	0.774	1.24	Top boosted tokens ↑
L22N1031-1💤 shadow specimen t0-v0.1	0.720	0.66	0.791	1.41	“ abandoning”“ avoided”“.lst”“�”“忘”“ avoidance”“慎用”“ reluctance”

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.