Layer 13 · Neuron 310 · suppressed

L13N310-1

Word quirks:

“ https”“ -”“ #”“⏎”“ Youtube”“ info”“ --”“ google”

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

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. I am designed to be an AI assistant that can perform complex tasks, such as answering questions, performing logical reasoning, and generating text based on content requirements (Qwen3.5) - free tier available on Coze Platform without registration. Coze: Coze-Platform: 1234567890:zh-CN.com.cn
</think>

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. L11N698-1🕒	0.77	“ https”“⏎”“<\|im_end\|>”
2. L2N2507+1💥	0.71	“ Whatsapp”“<\|im_end\|>”“ info”
3. L7N8909+1🕒	0.71	“ google”“ https”“ Youtube”
4. L4N5224+1🕒	0.71	“ google”“⏎”“ name”
5. L6N474-1💥	0.71	“<\|im_end\|>”“⏎”“ basically”
6. L3N3139-1🕒	0.70	“<\|im_end\|>”“⏎”“ Youtube”
7. L12N1895-1🕒	0.68	“<\|im_end\|>”“⏎”“ ?”
8. L0N103-1💥	0.68	“<\|im_end\|>”“ yahoo”“ youtube”
9. L3N4711-1🕒	0.68	“⏎”“ ⏎”“ info”
10. L14N1002+1💤	0.68	“ info”“ ⏎”“ goog”

Shadow specimen, apart from the ten: L13N310+1💤, 0.27.

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 310, 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
L13N310+1💤 shadow specimen t0-v0.1	0.770	0.68	0.958	1.36	“ ⏎”“ ⏎”“ ⏎⏎”“ ⏎”“⏎”“ ⏎”“ ⏎ ⏎”“ ***”
L13N310-1 t0-v0.1	1.295	0.56	1.550	1.49	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.