Layer 15 · Neuron 6490 · suppressed

L15N6490-1

Word quirks:

“ always”“ **“”“ sometimes”“ often”“ ALWAYS”“ your”“ frequently”“ “”

💤 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 Qwen3.5, the latest large language model developed by Tongyi Lab. My core capabilities include advanced text understanding, logical reasoning, and code generation. I can assist users with a wide range of tasks, from creative writing to technical problem-solving. How can I help you today? 😊

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. L17N1995-1💤	0.94	“ often”“ “”“ sometimes”
2. L15N8980+1🕒	0.89	“ always”“ sometimes”“ often”
3. L12N4809+1🕒	0.88	“ sometimes”“ always”“ usually”
4. L11N2767+1🕒	0.88	“ always”“ ALWAYS”“ sometimes”
5. L12N8503+1🕒	0.87	“ whenever”“ always”“ sometimes”
6. L11N3473-1🕒	0.87	“⏎”“ always”“ ALWAYS”
7. L10N2057-1💤	0.86	“ often”“ always”“ ALWAYS”
8. L11N681-1🕒	0.86	“ always”“ ⏎”“ often”
9. L12N1219-1🕒	0.85	“ always”“ now”“ sometimes”
10. L11N2219+1🕒	0.85	“ sometimes”“ your”“ always”

Shadow specimen, apart from the ten: L15N6490+1🕒, 0.20.

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 15, neuron 6490, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.43, top-1 agreement 0.77) 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
L15N6490+1🕒 shadow specimen t0-v0.1	2.070	0.48	2.428	2.10	“ only”“ ⏎”“ now”“ Only”“ ONLY”“ just”“";”“Only”
L15N6490-1 t0-v0.1	0.426	0.77	0.563	1.17	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.