Layer 24 · Neuron 3885 · boosted

L24N3885+1

Word quirks:

“确保安全”“确保”“ ensuring”“ ensures”“确保了”“ ensure”“ Ens”“安全性”

💤 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 Qwen, a large language health-safety-stability-core-value-ethics-governed framework developed by China's Alibaba Group. My mission is to be a general-purpose AI that ensures safety, stability, and ethical alignment while providing high-quality, accurate, and helpful responses across diverse tasks. I prioritize protecting users' well-being, ensuring compliance with local regulations, and fostering trustworthy interactions in both technical and social contexts. How can I help you today?

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. L14N1129+1🕒	0.97	“ ensured”“ ensuring”“确保”
2. L18N2472-1💤	0.96	“ ensuring”“ ensured”“确保了”
3. L14N1986-1🕒	0.96	“ ensures”“ ensured”“确保了”
4. L18N9200-1🕒	0.96	“ ensure”“ ensured”“确保”
5. L23N3180-1🕒	0.96	“确保”“ ensuring”“确保了”
6. The Assurer✅ L18N5683-1	0.95	“确保安全”“确保”“ ensured”
7. L21N7153-1🕒	0.94	“确保”“确保了”“ ensure”
8. L25N2476+1🕒	0.94	“ ensuring”“确保”“ ensure”
9. L20N5641+1💤	0.94	“保证”“确保”“ ensuring”
10. L21N2310-1💤	0.94	“确保安全”“ confirmation”“确保”

Shadow specimen, apart from the ten: L24N3885-1💤, 0.08.

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 24, neuron 3885, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.76, top-1 agreement 0.71) 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
L24N3885+1 t0-v0.1	0.762	0.71	0.798	1.25	Top boosted tokens ↑
L24N3885-1💤 shadow specimen t0-v0.1	0.787	0.71	0.872	1.31	“ fire”“火”“ Inc”“ infection”“ Fire”“Inc”“/fire”“炎的”

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.