Layer 10 · Neuron 3461 · suppressed

L10N3461-1

Word quirks:

“ typical”“ basically”“ typically”“ usually”“ generally”“ really”“ specific”“ seeming”

💤 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 Tongyi Lab. I am designed to assist you with your tasks by providing accurate, high-quality information in Chinese. My training data includes comprehensive knowledge up to my knowledge cutoff in 2026, which allows me to handle a wide range of topics and support complex requests effectively.

If you need help with specific questions or tasks, feel free to ask, and I will do my best to provide detailed, relevant, and helpful responses. Whether it's answering questions about the book or generating content, I aim to meet your needs with precision and care.

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. L11N5340+1💤	0.93	“ typically”“ generally”“ specifically”
2. L15N7539-1🕒	0.93	“ usually”“ typically”“ typical”
3. L11N1242+1💤	0.93	“ usually”“ typically”“ often”
4. L11N5706+1🕒	0.93	“ usually”“ typically”“ typical”
5. L10N6051+1💤	0.93	“ usually”“ typically”“ typical”
6. L8N4468-1🕒	0.92	“ usually”“ generally”“ typically”
7. L11N1934-1🕒	0.92	“ sometimes”“ usually”“ particular”
8. L9N819+1🕒	0.92	“ usually”“ specific”“ typically”
9. L9N1256+1💤	0.92	“ typically”“ specific”“ usually”
10. L9N5897-1🕒	0.92	“ typical”“ normally”“ usually”

Shadow specimen, apart from the ten: L10N3461+1🕒, -0.01.

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 10, neuron 3461, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.92, top-1 agreement 0.65) 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
L10N3461+1🕒 shadow specimen t0-v0.1	1.160	0.57	1.407	1.47	“ ⏎”“ ⏎⏎”“ ”“ here”“...”“...””“ ____”“<\|im_end\|>”
L10N3461-1 t0-v0.1	0.919	0.65	1.066	1.03	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.