Layer 24 · Neuron 3229 · suppressed

L24N3229-1

Word quirks:

“ horizontal”“Horizontal”“ Horizontal”“ Hor”“-horizontal”“horizontal”“ vertical”“-vertical”

💤 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 Q90, a large-scale language model developed by the world's largest AI company. My mission is to provide high-quality, accurate, and safe responses to your questions across a wide range of topics, while ensuring transparency and ethical integrity in all interactions. I aim to be your reliable partner in learning, problem-solving, and creative tasks, always prioritizing clarity and trust. 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. L24N4777+1💥	0.97	“ horizontal”“ Horizontal”“水平”
2. L5N5918-1🕒	0.97	“ horizontal”“ vertical”“ horizontally”
3. L26N5760+1🕒	0.97	“ horizontal”“ Horizontal”“horizontal”
4. L0N7035+1💥	0.96	“<\|im_end\|>”“ horiz”“vertical”
5. L25N2901-1🕒	0.96	“ horizontal”“ Horizontal”“横向”
6. L30N6343+1🕒	0.95	“horizontal”“ horizontal”“ Horizontal”
7. L28N6684-1🕒	0.95	“Horizontal”“ horizontal”“ horizontally”
8. L26N5109+1🕒	0.95	“-horizontal”“horizontal”“(horizontal”
9. L20N2673-1🕒	0.95	“ horizontal”“ vertical”“的水平”
10. L2N3567+1💥	0.94	“<\|im_end\|>”“ horizontal”“ horiz”

Shadow specimen, apart from the ten: L24N3229+1🕒, -0.11.

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 24, neuron 3229, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.70, top-1 agreement 0.73) 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
L24N3229+1🕒 shadow specimen t0-v0.1	0.863	0.72	0.881	1.34	“ enzym”“恩”“酶”“ En”“ enzyme”“ enzymes”“氧化”“En”
L24N3229-1 t0-v0.1	0.703	0.73	0.758	1.22	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.