Layer 23 · Neuron 8591 · suppressed

L23N8591-1

Word quirks:

“芹菜”“三元”“芹”“阵”“宽”“тет”“钰”“小草”

💤 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 CMMT, a specialized AI model designed to assist in the field of computer science and mathematics. My core mission is to serve as a reliable, high-precision partner for professionals and researchers in these disciplines.

My specific objectives include:

Code Generation & Analysis: Writing, debugging, and explaining complex code snippets across multiple programming languages.
Mathematical Problem Solving: T tackling rigorous mathematical problems, proofs, and statistical analysis with step-by-step reasoning.
System Design: Helping architects build scalable systems, algorithms, and data structures.
**Knowledge Retrie

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. L30N5475+1💤	0.86	“形”“参”“名”
2. L31N5488+1💥	0.86	“胡”“黄”“柳”
3. L30N9136-1🕒	0.86	“采”“萧”“鼎”
4. L28N6339+1💤	0.86	“太”“芮”“玉”
5. L28N4146-1💤	0.86	“吹”“鼎”“青”
6. L21N4615+1💤	0.85	“薯”“舍”“ $$$”
7. L26N8298-1💤	0.85	“仕”“商”“坚毅”
8. L26N9142-1🕒	0.85	“阵”“陣”“股”
9. L26N3751-1🕒	0.85	“牛”“绵”“茅”
10. L31N7989+1💥	0.85	“愁”“堂”“章”

Shadow specimen, apart from the ten: L23N8591+1💤, 0.48.

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 23, neuron 8591, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.72, top-1 agreement 0.71) 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
L23N8591+1💤 shadow specimen t0-v0.1	0.721	0.75	0.707	1.23	“核”“ pits”“ plum”“ кост”“ kernel”“ pit”“ rais”“浆”
L23N8591-1 t0-v0.1	0.725	0.71	0.837	1.35	Top boosted tokens ↑

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.