Layer 11 · Neuron 4990 · boosted

L11N4990+1

Word quirks:

“ sometimes”“ usually”“ now”“ today”“ scenario”“ often”“ here”“ always”

💤 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, a large language model developed by Alibaba Cloud. My mission is to assist users with a wide range of tasks, such as answering questions, generating text, or providing information. I can help with anything from simple queries to complex problem-solving, always aiming to be helpful and informative. How can I assist 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. L10N3046-1💤	0.92	“ here”“ now”“ often”
2. L15N1220+1💤	0.91	“ often”“ usually”“ sometimes”
3. L13N7851-1🕒	0.91	“ usually”“ sometimes”“ similar”
4. L10N8727+1🕒	0.90	“ often”“ sometimes”“ today”
5. L11N3641-1🕒	0.89	“ usually”“ now”“ sometimes”
6. L11N1026-1🕒	0.89	“ often”“ sometimes”“ usually”
7. L12N695-1💤	0.89	“ sometimes”“ often”“ more”
8. L10N4749-1💤	0.89	“ often”“ sometimes”“ typically”
9. L17N6283+1💤	0.89	“ often”“ always”“ today”
10. L13N2215-1🕒	0.88	“ always”“ sometimes”“ often”

Shadow specimen, apart from the ten: L11N4990-1🕒, 0.35.

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 11, neuron 4990, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.71, top-1 agreement 0.68) 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
L11N4990+1 t0-v0.1	0.714	0.68	0.962	1.36	Top boosted tokens ↑
L11N4990-1🕒 shadow specimen t0-v0.1	1.082	0.61	1.338	1.45	“ continued”“<\|im_end\|>”“ continuation”“ continuing”“ now”“ continues”“ continue”“ ⏎⏎”

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.