Layer 21 · Neuron 2310 · suppressed

L21N2310-1

Word quirks:

“确保安全”“ confirmation”“确保”“确保了”“确认”“ ensuring”“保障”“Ensure”

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

Identity (according to the specimen itself)

Hello! I am Q*, the latest model in the *Q series of large language models developed by Q (the company). My mission is to provide high-quality assistance across a wide range of tasks, from answering questions and creating content to reasoning complex problems and ensuring safety and ethical compliance. Whether it's helping with creative writing, coding, or just having a conversation, I'm here to offer thoughtful and accurate support. What would you like to explore together? 😊

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. The Assurer✅ L18N5683-1	0.95	“确保安全”“确保”“ ensured”
2. L24N3885+1💤	0.94	“确保安全”“确保”“ ensuring”
3. L25N923+1🕒	0.94	“确保”“保证”“ ensuring”
4. L18N9200-1🕒	0.94	“ ensure”“ ensured”“确保”
5. L23N1551-1🕒	0.93	“确保”“保证”“ ensures”
6. L20N5641+1💤	0.93	“保证”“确保”“ ensuring”
7. L23N3180-1🕒	0.93	“确保”“ ensuring”“确保了”
8. L24N5376+1💤	0.93	“提出”“确保”“ ensured”
9. L18N2472-1💤	0.91	“ ensuring”“ ensured”“确保了”
10. L21N4903+1🕒	0.91	“确保”“ ensuring”“并确保”

Shadow specimen, apart from the ten: L21N2310+1💤, 0.06.

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 21, neuron 2310, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.66) 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
L21N2310+1💤 shadow specimen t0-v0.1	0.720	0.71	0.789	1.15	“五项”“ Three”“ Six”“Six”“ Four”“ Five”“四项”“ identities”
L21N2310-1 t0-v0.1	0.728	0.66	0.896	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.