Layer 21 · Neuron 2794 · boosted

L21N2794+1

Word quirks:

“表面”“ surface”“Surface”“ Surface”“ surfaces”“表面的”“支撑”“surface”

💤 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 Q, a research-tunable support desk for the Q platform. My mission is to provide clear, concise, and actionable information by summarizing key details from official sources while ensuring accuracy and relevance to your needs. I do not generate creative content, personal narratives, or speculative ideas—my focus is on delivering precise, reliable support through structured responses. If you have a specific question or need a summary of a topic, I can help! 📊✨

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. L26N2949-1🕒	0.98	“表面”“ surfaces”“ surface”
2. L25N4158-1🕒	0.98	“表面”“ Surface”“ surface”
3. L25N7134-1🕒	0.98	“ surface”“表面”“ Surface”
4. L22N6907+1🕒	0.98	“ surface”“ surfaces”“ Surface”
5. L31N1190-1💥	0.98	“ surface”“表面”“ Surface”
6. L30N6952+1🕒	0.97	“表面”“ surface”“表面的”
7. L26N3368+1🕒	0.97	“ surface”“ Surface”“Surface”
8. L24N5797-1🕒	0.97	“ surface”“ Surface”“surface”
9. L25N3542+1🕒	0.97	“ surface”“表面”“ Surface”
10. L19N4527+1💤	0.97	“ surface”“表面”“ Surface”

Shadow specimen, apart from the ten: L21N2794-1💤, 0.14.

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 21, neuron 2794, 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 0.68) 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
L21N2794+1 t0-v0.1	0.748	0.68	0.838	1.30	Top boosted tokens ↑
L21N2794-1💤 shadow specimen t0-v0.1	0.610	0.72	0.698	1.26	“纠”“_given”“ similar”“ indire”“ hinter”“similar”“複雜”“ globally”

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.