Layer 17 · Neuron 1117 · suppressed

L17N1117-1

Word quirks:

“在”“，”“直接”“为”“是”“具体”“的”“、”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

我是Qwen3.5，是阿里巴巴大模型在人类反馈的强化学习（RLHF）方面进行了优化。我的使命是提供高质量、安全、高效的智能支持，具体体现在：

精准理解与表达：通过多轮对话和上下文理解，确保交流自然流畅。
任务执行：无论是代码生成、逻辑推理、多步骤规划，还是创意写作，我 strive to deliver actionable results.
跨领域应用：从医疗、法律到编程、艺术创作，我 aim to support diverse应用场景.
**安全

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. L11N135+1💥	0.97	“为”“，”“在”
2. L10N7844+1💥	0.95	“在”“与”“，”
3. L11N3415+1💥	0.95	“在”“，”“与”
4. L2N2253-1💥	0.95	“，”“在”“的”
5. L6N820+1💥	0.95	“在”“。”“与”
6. L14N3516+1🕒	0.95	“（”“直接”“，”
7. L12N2489+1💥	0.95	“在”“的”“，”
8. L18N3960+1💥	0.95	“在”“的”“，”
9. L6N146+1💥	0.95	“是”“的”“为”
10. L10N7316-1💥	0.95	“。”“，”“在”

Shadow specimen, apart from the ten: L17N1117+1🕒, 0.40.

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 17, neuron 1117, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.71, top-1 agreement 0.49) 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
L17N1117+1🕒 shadow specimen t0-v0.1	2.055	0.44	2.378	2.86	“ được”“ ở”“ có”“ không”“ đã”“ một”“ trong”“ các”
L17N1117-1 t0-v0.1	1.715	0.49	1.836	1.52	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.