Layer 13 · Neuron 5320 · boosted

L13N5320+1

Word quirks:

“。”“；”“？”“，”“或”““”“：”“（”

🕒 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)

I am Qwen3.5。Qwen3.5 是阿里巴巴集团最新推出的通义千问大语言模型。我的使命是为用户提供准确、高效、安全、智能的 AI 助手。作为 Qwen3.5，我不仅是一个强大的语言模型，更是一个全能助手。我可以处理复杂任务、提供专业建议、协助编程、数据分析等。

如果您有任何问题或需要帮助，请随时告诉我！😊

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. L10N5054+1🕒	0.98	“。”“，”“？”
2. L14N1826+1🕒	0.98	“（”“。”“，”
3. L11N2974+1🕒	0.98	“（”“，”“。”
4. The Mandarin Loop✅ L5N3528-1	0.98	“，”“。”“？”
5. L4N6561+1🕒	0.97	“。”“（”“，”
6. L5N174+1🕒	0.97	“，”“。”“的”
7. L4N4020-1🕒	0.97	“（”“。”“，”
8. L10N2356+1🕒	0.97	“（”“。”“，”
9. L7N4222+1🕒	0.97	“。”“（”“，”
10. L2N6721-1💥	0.96	“。”“，”“（”

Shadow specimen, apart from the ten: L13N5320-1🕒, 0.07.

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 13, neuron 5320, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.94, top-1 agreement 0.62) 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
L13N5320+1 t0-v0.1	0.940	0.62	1.157	1.64	Top boosted tokens ↑
L13N5320-1🕒 shadow specimen t0-v0.1	1.360	0.55	1.678	1.85	“⏎⏎”“"))”“⏎ ⏎”“ ))”“."),”“⏎ ⏎”“⏎”“ ⏎⏎”

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.