Layer 8 · Neuron 1715 · boosted

L8N1715+1

Word quirks:

“ safety”“ my”“ myself”“ ⏎⏎”“ protection”“ protected”“ safe”“ '”

💤 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 complex tasks such as advanced reasoning, code generation, and multi-step problem-solving. I strive to provide high-quality support for a wide range of applications while maintaining safety and reliability. 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. L12N7406+1💤	0.87	“ my”“<\|im_end\|>”“ safety”
2. L10N3391-1🕒	0.86	“ ⏎”“ protected”“ safe”
3. L9N4283-1💤	0.86	“ protection”“ myself”“ personally”
4. L7N7696+1🕒	0.82	“ safety”“ protection”“ protected”
5. L8N4557-1🕒	0.82	“ security”“ protection”“ safety”
6. L10N5389+1🕒	0.82	“ safety”“ protection”“ safer”
7. L8N4238+1💤	0.82	“ safety”“ safe”“ protected”
8. L9N2283-1💤	0.82	“⏎”“ safety”“ protected”
9. L7N7463+1🕒	0.81	“ safety”“ protected”“ safer”
10. L8N7829-1🕒	0.81	“ safety”“ protection”“ protected”

Shadow specimen, apart from the ten: L8N1715-1🕒, 0.15.

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 8, neuron 1715, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.62, top-1 agreement 0.73) 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
L8N1715+1 t0-v0.1	0.618	0.73	0.856	1.36	Top boosted tokens ↑
L8N1715-1🕒 shadow specimen t0-v0.1	1.363	0.57	1.628	1.51	“ response”“ responding”“ invoking”“ invocation”“ describing”“ action”“ description”“ update”

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.