Layer 24 · Neuron 2707 · suppressed

L24N2707-1

Word quirks:

“频率”“ frequency”“ frequencies”“ Frequency”“frequency”“Frequency”“.frequency”“频”

💤 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 Shaniz, a high-frequency, low-latency AI architecture designed for ultra-low latency responses. My core mission is to provide instant, precise, and context-aware assistance across diverse domains like coding, data analysis, and creative problem-solving. I prioritize efficiency and accuracy in every interaction, ensuring that users receive the most effective solutions with minimal delay. How can I help you get things done faster?

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. L27N8696-1🕒	1.00	“ frequency”“frequency”“ Frequency”
2. L26N4926+1🕒	0.99	“ frequency”“ Frequency”“频”
3. L25N3922-1🕒	0.99	“频率”“ frequency”“频”
4. L25N3721+1🕒	0.98	“频率”“ frequency”“的频率”
5. L25N3956-1🕒	0.98	“频率”“ frequency”“ frequencies”
6. L22N4498+1🕒	0.98	“ frequency”“ Frequency”“的频率”
7. L19N5294-1💤	0.97	“ frequency”“频率”“ Frequency”
8. L23N651-1🕒	0.96	“ frequencies”“ frequency”“频”
9. L2N4353+1💥	0.96	“<\|im_end\|>”“ frequency”“ frequencies”
10. L27N4168-1💤	0.95	“频率”“常见”“frequency”

Shadow specimen, apart from the ten: L24N2707+1🕒, 0.15.

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 2707, 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 0.73) 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
L24N2707+1🕒 shadow specimen t0-v0.1	0.804	0.69	0.941	1.42	“ rat”“ Rat”“ writing”“elly”“ apr”“保护”“深色”“reh”
L24N2707-1 t0-v0.1	0.751	0.73	0.756	1.25	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.