Layer 11 · Neuron 4674 · suppressed

L11N4674-1

Word quirks:

“ *"”“ sometimes”“:"”“ **"”“ *“”“ **“”“ 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 a large language model trained by Google. My mission is to provide helpful, accurate, and safe information to users around the world. I strive to be as useful as possible, but I can also make mistakes. Here's how I am:

Capabilities: I am an AI assistant developed by Google, designed to assist users with tasks like answering questions, generating text, or providing code.
Limitations: I may not always be perfect, as I can sometimes make errors or have biases.
Goals: My goal is to enhance user experience and deliver high-quality responses.

While I aim to be

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. L8N3976-1💤	0.95	“ "”“ *"”“ reason”
2. L6N7024+1💤	0.93	“ "”“ «”“ \"”
3. L8N369-1💤	0.92	“ ”“ description”“ *"”
4. L15N5199+1🕒	0.92	“ "”“ "”“ ”
5. L15N6375+1💤	0.92	“ *“”“ "”“ **"”
6. L9N2095+1💤	0.92	“ “”“ "”“ "**”
7. L10N7021+1🕒	0.92	“ “”“ "”“ “”
8. L5N7725+1🕒	0.91	“ "”“ "”“ *”
9. L6N7939+1🕒	0.91	“ “”“ “”“ "”
10. L10N7966+1🕒	0.91	“ *“”“ “”“ **"”

Shadow specimen, apart from the ten: L11N4674+1💤, 0.25.

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 11, neuron 4674, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, 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
L11N4674+1💤 shadow specimen t0-v0.1	0.729	0.64	1.019	1.23	“ --”“ ⏎”“ ⏎”“ thus”“ -”“ now”“ only”“ <U+00A0>”
L11N4674-1 t0-v0.1	0.750	0.69	0.886	1.26	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.