Layer 20 · Neuron 2892 · boosted

L20N2892+1

Word quirks:

“选择题”“选好”“ choice”“ secured”“ اخت”“ thinner”“选一”“ choosing”

💤 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 Q3.5, a large language model developed by Tencent DeepLabFM. My mission is to assist users with various tasks, including answering questions, solving problems, creating content, and more. I aim to provide accurate, creative, and helpful responses while ensuring safety and ethical considerations. Let me know if you'd like to try something specific! 🚨👋

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. L22N3893+1🕒	0.87	“选择的”“选择一个”“ chosen”
2. L22N6647+1🕒	0.87	“选择”“ choice”“ selection”
3. L20N3104+1🕒	0.87	“自由选择”“ choosing”“选择了”
4. L27N6455+1🕒	0.86	“选”“选择”“ choices”
5. L21N5954-1💤	0.86	“ choice”“ chosen”“的选择”
6. L21N2228-1💤	0.86	“ nicely”“选择和”“ choose”
7. L23N5547-1🕒	0.86	“选择”“ choice”“ choosing”
8. L19N7741+1🕒	0.86	“ choose”“选择”“ choosing”
9. L21N4322+1💤	0.85	“选择”“ chosen”“选择的”
10. L31N768-1💥	0.85	“选择”“ choice”“ choose”

Shadow specimen, apart from the ten: L20N2892-1🕒, 0.21.

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 20, neuron 2892, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.60, top-1 agreement 0.72) 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
L20N2892+1 t0-v0.1	0.600	0.72	0.723	1.16	Top boosted tokens ↑
L20N2892-1🕒 shadow specimen t0-v0.1	1.022	0.60	1.140	1.32	“ BEST”“ MOST”“ best”“ Best”“ most”“best”“Best”“最直接”

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.