Layer 9 · Neuron 2466 · boosted

L9N2466+1

Word quirks:

“ \"”“ sometimes”“ related”“ "*”“ reflected”“ *"”“ info”“ description”

💤 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 provide high-quality information, answer questions, and assist users in various tasks related to technology, science, and culture. I aim to be a helpful, harmless, and honest AI assistant, contributing to the advancement of knowledge and fostering positive interactions. Whether it's discussing the latest advancements in artificial intelligence or exploring other topics, I strive to support users in achieving their goals.

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. L9N6100+1🕒	0.77	“ *“”“ \"”“ related”
2. L12N793-1🕒	0.75	“ stated”“ sometimes”“ *"”
3. L11N2811-1💤	0.74	“ "”“ "”“ always”
4. L12N557-1💤	0.73	“ *“”“ description”“ info”
5. L11N7171-1🕒	0.70	“ often”“ "”“ "”
6. Spinout✅ L12N8445-1	0.70	“ info”“<\|im_end\|>”“ \"”
7. L9N5826-1💤	0.69	“ *"”“ related”“ "”
8. L9N3549+1💤	0.69	“ *"”“ specific”“ ”
9. L10N6712-1🕒	0.69	“ "”“ "”“ *"”
10. L12N8808-1🕒	0.68	“ "”“ "*”“ "”

Shadow specimen, apart from the ten: L9N2466-1🕒, 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 9, neuron 2466, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.85, top-1 agreement 0.62) 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
L9N2466+1 t0-v0.1	0.854	0.62	1.091	1.44	Top boosted tokens ↑
L9N2466-1🕒 shadow specimen t0-v0.1	1.307	0.55	1.574	1.47	“ number”“ multiple”“ count”“ numbers”“-number”“ typically”“ meaning”“ numbering”

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.