Layer 13 · Neuron 8581 · suppressed

L13N8581-1

Word quirks:

“ adequately”“⏎”“ insufficient”“ ⏎”“ suitable”“ sufficiently”“ desired”“ fit”

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

Identity (according to the specimen itself)

I'm Qwen3.5, a large language model developed by Alibaba's Tongyi Lab. My capabilities include generating high-quality text across a wide range of tasks: writing, reasoning, coding, and adhering to specific style or length requirements. I can also handle complex scenarios like summarizing, translating, or following precise instructions in terms of accuracy and tone.

To better assist you, I may need to adjust my approach based on your specific needs. If you'd like, I can help you with:

Role: Generating content that aligns with your goals (e.g., creative writing, technical documentation, etc

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. L15N6083+1🕒	0.87	“ typical”“ suitable”“ enough”
2. L15N8445+1🕒	0.85	“ lacks”“ sufficiently”“ adequate”
3. L14N7588+1🕒	0.83	“ suitable”“ appropriately”“ acceptable”
4. L18N8007+1💤	0.83	“ sufficient”“ satisfactory”“ enough”
5. L14N7447-1💤	0.82	“ enough”“ needed”“ straightforward”
6. L16N5579+1🕒	0.82	“ adequately”“ appropriately”“ clearly”
7. L17N7636+1💤	0.82	“⏎”“ reasonable”“ appropriate”
8. L17N4177-1🕒	0.81	“ sufficient”“ sufficiently”“ enough”
9. L15N7323-1🕒	0.81	“ suitable”“ appropriate”“ this”
10. L14N8792+1🕒	0.81	“ enough”“ sufficiently”“ sufficient”

Shadow specimen, apart from the ten: L13N8581+1🕒, 0.02.

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 8581, 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
L13N8581+1🕒 shadow specimen t0-v0.1	0.839	0.66	1.000	1.07	“ hidden”“ hiding”“隐藏”“ obscure”“ concealed”“ Hidden”“ mysterious”“ mystery”
L13N8581-1 t0-v0.1	1.218	0.60	1.456	1.53	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.