Layer 15 · Neuron 6905 · boosted

L15N6905+1

Word quirks:

“ ***”“ __”“ *”“ **”“ —”“ *\”“ _____”“ ____”

💤 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 provide high-quality assistance across various domains, including answering questions, generating content, and solving complex problems. I strive to be helpful, accurate, and ethical in my interactions, ensuring that users receive reliable information and guidance.

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. L13N8113-1🕒	0.97	“ ”“ ____”“ **”
2. L7N3911-1💤	0.94	“ ____”“ ______”“ ***”
3. L10N6444+1🕒	0.94	“ *”“ ”“ ______”
4. L12N8612-1🕒	0.94	“ **”“ ____”“ ”
5. L13N361-1💤	0.94	“ ”“ ”“ ”
6. L14N9008-1💤	0.93	“ ”“ [”“ **”
7. L10N7437-1💤	0.93	“ ”“ "”“ *”
8. L10N4546-1🕒	0.93	“ ____”“ __”“ --”
9. L9N1844-1🕒	0.93	“ ***”“ _____”“ ____”
10. L14N8480-1💤	0.93	“ ”“ **”“ ____”

Shadow specimen, apart from the ten: L15N6905-1🕒, -0.03.

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 6905, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.57, top-1 agreement 0.67) 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
L15N6905+1 t0-v0.1	0.569	0.67	0.772	1.15	Top boosted tokens ↑
L15N6905-1🕒 shadow specimen t0-v0.1	0.748	0.69	1.024	1.47	“ now”“ new”“ ours”“ your”“ yours”“ does”“ needed”“ doesn”

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.