Layer 4 · Neuron 8641 · suppressed

L4N8641-1

Word quirks:

“ stating”“ saying”“ said”“ statements”“ mentioning”“ statement”“ referring”“ comment”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen, a large language model developed by Alibaba Cloud. My mission is to assist users in various tasks such as answering questions, performing actions, and providing insights. 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. L14N5832+1🕒	0.92	“ stating”“ statement”“ stated”
2. L15N3299-1🕒	0.92	“ info”“ saying”“ stating”
3. L6N1804+1🕒	0.91	“ saying”“ say”“ said”
4. L17N3118-1🕒	0.91	“ stating”“ saying”“ stated”
5. L14N1959+1🕒	0.89	“ stating”“ mentioning”“ your”
6. L5N3202-1💥	0.89	“ stating”“ saying”“ say”
7. L10N4885+1🕒	0.88	“ saying”“ stating”“ expressing”
8. L4N3543-1🕒	0.87	“ said”“ stated”“ stating”
9. L19N4798+1🕒	0.87	“ stated”“ statement”“ mentioned”
10. L14N2478+1💤	0.86	“ statement”“ statements”“ claim”

Shadow specimen, apart from the ten: L4N8641+1🕒, 0.22.

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 4, neuron 8641, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.91, top-1 agreement 0.42) 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
L4N8641+1🕒 shadow specimen t0-v0.1	1.091	0.64	1.304	1.80	“⏎”“!”“ ‘”“?!”“ *“”“ ”“ himself”“ **"”
L4N8641-1 t0-v0.1	1.914	0.42	2.286	2.15	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.