Layer 20 · Neuron 8430 · boosted

L20N8430+1

Word quirks:

“用户体验”“ experience”“体验”“ Experience”“ experiences”“Experience”“experience”“體驗”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Qingwen, an advanced large-scale language model developed with a focus on providing high-quality user experiences. My mission is to deliver exceptional interactions by ensuring strong user satisfaction, prioritizing user experience, and guaranteeing high quality. I strive to provide the best possible answers and support for your queries, whether they involve improving user experience, enhancing quality of life, or delivering top-tier service.

If you have any questions or need assistance, feel free to ask!

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. L29N8442-1💤	0.97	“ Experience”“ experiences”“的体验”
2. L25N9203+1🕒	0.96	“体验”“ experience”“ Experience”
3. L24N4863-1🕒	0.96	“ experiences”“ experience”“ Experience”
4. L28N2608+1🕒	0.96	“体验”“ experience”“经验”
5. L21N8692+1🕒	0.95	“ experiences”“ experience”“ experiencing”
6. L31N1148-1💥	0.95	“ experience”“ Experience”“experience”
7. L31N6192+1💥	0.94	“ experience”“ Experience”“experience”
8. L22N8734+1🕒	0.94	“ experiences”“ experience”“ Experience”
9. L20N5511+1🕒	0.94	“ experience”“ experiences”“experience”
10. L25N1760+1🕒	0.93	“ experience”“experience”“ experiences”

Shadow specimen, apart from the ten: L20N8430-1🕒, 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 20, neuron 8430, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L20N8430+1 t0-v0.1	1.235	0.62	1.410	1.44	Top boosted tokens ↑
L20N8430-1🕒 shadow specimen t0-v0.1	0.665	0.71	0.768	1.22	“救灾”“ cor”“免”“ prem”“ovens”“ Power”“救治”“ healed”

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.