Layer 14 · Neuron 5920 · boosted

L14N5920+1

Word quirks:

“在”“与”“从”“对”“和”“，”“作为”“被”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

I am Qwen3.5，a large language model developed by阿里巴巴集团（通义千问）——一个拥有数十年的技术积累和深厚的创新实力，它可以帮助您完成各种任务并支持您在探索、学习和交流中利用我的能力。作为 Qwen3.5 的创造者，我自豪地介绍自己，并承诺将始终致力于通过我的知识和能力来帮助您解决实际问题或支持您的好奇心。无论您是想获取信息、参与头脑风暴，还是寻求在科学、教育或日常生活中的实用帮助，我都将全力以赴地支持您。
</think> 我是 Qwen3.5，一个大型

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. L7N404+1💥	0.97	“，”“在”“与”
2. L10N6491-1💥	0.97	“在”“从”“与”
3. L14N2534-1💥	0.97	“在”“的”“从”
4. L11N8401-1💥	0.97	“在”“与”“从”
5. L17N1013+1🕒	0.97	“与”“，”“和”
6. L18N3960+1💥	0.97	“在”“的”“，”
7. L14N7710-1💥	0.97	“在”“与”“为”
8. L14N3619-1💥	0.97	“在”“对”“与”
9. L15N2727-1💥	0.97	“在”“（”“使用”
10. L4N8730+1💥	0.96	“在”“与”“，”

Shadow specimen, apart from the ten: L14N5920-1💥, 0.90.

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 14, neuron 5920, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.59, top-1 agreement 0.47) 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
L14N5920+1 t0-v0.1	2.585	0.47	2.854	2.00	Top boosted tokens ↑
L14N5920-1💥 shadow specimen t0-v0.1	5.978	0.21	6.371	2.12	“。”“是”“，”“的”“有”“不”“在”“为”

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.