Layer 9 · Neuron 6212 · suppressed

L9N6212-1

Word quirks:

“ displayed”“ visually”“ width”“ pictures”“ display”“ visuals”“ images”“ usage”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

Hello! I am Qwen3.5, a large language model developed by Alibaba Cloud. My name is Qwen3.5, and I am designed to assist users with a wide range of tasks.

Who am I?

I am an AI assistant with the following key features:

High Intelligence: I possess strong reasoning capabilities and can handle complex problems.
Multilingual Support: I support multiple languages including Chinese, English, French, Spanish, Portuguese, and others.
Visual Understanding: I can interpret images, charts, and diagrams to provide accurate information.
**Code Generation

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. L5N4173+1🕒	0.92	“ display”“ images”“ camera”
2. L10N8532+1🕒	0.91	“ image”“ visual”“ visible”
3. L7N5204-1🕒	0.90	“ visual”“ display”“ displaying”
4. L8N5167-1🕒	0.90	“ image”“ visual”“ overlay”
5. L9N8050-1🕒	0.90	“ visual”“ display”“ visually”
6. L6N8776-1🕒	0.90	“ display”“ displays”“ displayed”
7. L10N5387-1🕒	0.89	“ visual”“ images”“ visually”
8. The Phantom Image✅ L5N4989-1	0.88	“ visual”“ display”“ math”
9. L9N6122+1🕒	0.87	“ visual”“ display”“ visuals”
10. L9N3273-1🕒	0.86	“ image”“ visual”“ visually”

Shadow specimen, apart from the ten: L9N6212+1🕒, 0.05.

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 6212, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.64) 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
L9N6212+1🕒 shadow specimen t0-v0.1	1.512	0.54	1.827	1.98	“ discussion”“<\|im_end\|>”“ acknowledgment”“ response”“ referenced”“ discussions”“ communication”“ reference”
L9N6212-1 t0-v0.1	0.856	0.64	1.092	1.19	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.