Layer 15 · Neuron 1397 · suppressed

L15N1397-1

Word quirks:

“?"”“？””“?\”“?””“"?”“?'”“?</”“?**”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model trained by Alibaba Cloud. My mission is to assist users in various tasks, including answering questions, generating content, and providing creative solutions. I aim to be helpful, harmless, and honest, ensuring that my responses are accurate and relevant to the user's needs.

---

Qwen3.5
My mission: To assist users in various tasks, including answering questions, generating content, and providing creative solutions. I aim to be helpful, harmless, and honest, ensuring that my responses are accurate and relevant to the user's needs."

---

**Q

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. L14N6918-1💤	0.99	“?**”“?\”“?””
2. L2N1831+1💥	0.99	“?\”“?"”“?</”
3. L1N3016+1💥	0.99	“?"”“?\”“"?”
4. L6N4737+1🕒	0.99	“?\”“?”“？”
5. L4N4424+1💥	0.99	“?\”“?"”“?</”
6. L6N4497-1🕒	0.99	“?\”“?"”“?”
7. L10N3218+1🕒	0.98	“?\”“?"”“?””
8. L13N1874+1🕒	0.98	“？””“?"”“?””
9. L3N1801-1🕒	0.98	“?\”“<\|im_end\|>”“?"”
10. L9N123-1🕒	0.98	“?\”“?”“?</”

Shadow specimen, apart from the ten: L15N1397+1🕒, -0.00.

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 15, neuron 1397, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.50, top-1 agreement 0.72) 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
L15N1397+1🕒 shadow specimen t0-v0.1	0.783	0.63	0.978	1.30	“⏎”“ ⏎”“ ⏎”“ ⏎”“ ⏎”“⏎ ⏎”“ ⏎”“⏎ ⏎”
L15N1397-1 t0-v0.1	0.504	0.72	0.652	1.27	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.