Layer 15 · Neuron 5798 · suppressed

L15N5798-1

Word quirks:

“ sometimes”“ previously”“ clearer”“ often”“ clear”“ Sometimes”“sometimes”“ may”

💤 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's Tongyi Lab. My mission is to assist users in achieving their goals through high-quality interaction and problem-solving. Whether it's creative writing, logical reasoning, or complex coding tasks—I strive to be as helpful, accurate, and efficient as possible. I also support long-context understanding and multi-modal processing to ensure precision in my responses. If you have any specific questions or tasks you need help with, and I will do my best to provide clear, actionable 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. L14N4792-1💤	0.86	“ sometimes”“ maybe”“ often”
2. L14N7605-1🕒	0.86	“ sometimes”“ maybe”“ occasionally”
3. L13N9113+1🕒	0.86	“ sometimes”“ often”“ occasionally”
4. L14N6924-1🕒	0.85	“ sometimes”“ possibly”“ some”
5. L16N3903-1🕒	0.85	“ sometimes”“ occasionally”“ Sometimes”
6. L16N6395+1🕒	0.85	“ sometimes”“ Sometimes”“ carefully”
7. L19N1522-1💤	0.85	“ sometimes”“ usually”“ possibly”
8. L13N5622-1🕒	0.84	“ sometimes”“ somewhat”“ Sometimes”
9. L15N1641-1💤	0.84	“ sometimes”“ often”“ always”
10. L13N5687+1🕒	0.84	“ sometimes”“sometimes”“ occasionally”

Shadow specimen, apart from the ten: L15N5798+1🕒, 0.24.

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 5798, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.59, top-1 agreement 0.69) 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
L15N5798+1🕒 shadow specimen t0-v0.1	0.907	0.67	1.161	1.25	“ see”“ feel”“ experience”“ look”“ avoid”“ try”“ similar”“ ⏎”
L15N5798-1 t0-v0.1	0.585	0.69	0.738	1.30	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.