Layer 12 · Neuron 5235 · suppressed

L12N5235-1

Word quirks:

“<|im_end|>”“?\”“ ____”“ ______”“ seems”“ seemed”“ ???”“？**”

🕒 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 Qwen3.5, a large language model developed by Alibaba. My mission is to help users with various tasks, such as answering questions, writing stories, or solving math problems. I strive to provide accurate information and avoid any errors in my responses. I also try to be helpful and friendly in my interactions, even when discussing complex topics.

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. L11N6526-1🕒	0.88	“?”“?\”“ ____”
2. L5N3795-1💥	0.85	“<\|im_end\|>”“？”“?\”
3. L14N2949-1🕒	0.83	“ seems”“?\”“ seem”
4. L6N4573+1🕒	0.82	“ ____”“ seems”“ ???”
5. L15N932+1🕒	0.82	“<\|im_end\|>”“ seems”“ ____”
6. L12N4175+1🕒	0.81	“?\”“\n”“ will”
7. L1N8412+1💥	0.80	“<\|im_end\|>”“ _____”“?\”
8. L3N4271+1🕒	0.80	“<\|im_end\|>”“？”“?\”
9. L3N1033+1🕒	0.80	“?”“?\”“?"”
10. L1N5437-1💥	0.80	“<\|im_end\|>”“ ?”“ ____”

Shadow specimen, apart from the ten: L12N5235+1🕒, -0.07.

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 12, neuron 5235, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.95, top-1 agreement 0.66) 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
L12N5235+1🕒 shadow specimen t0-v0.1	1.008	0.61	1.337	1.60	“ momentarily”“ incorrectly”“ temporarily”“ needed”“ currently”“ already”“ awaiting”“ misrepresented”
L12N5235-1 t0-v0.1	0.949	0.66	1.201	1.45	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.