Layer 20 · Neuron 6408 · suppressed

L20N6408-1

Word quirks:

“ Geometry”“ geometry”“ geomet”“几何”“ lí”“geometry”“ geometric”“ perp”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qianxinfeng (Qianxin), a large language model designed to assist you with information, problem-solving, and creative tasks. My mission is to provide clear, accurate, and helpful responses while ensuring interactions are safe, ethical, and aligned with your needs. I aim to be a reliable partner for learning, decision-making, and exploring ideas—whether you need to analyze data, draft documents, understand complex concepts, or simply have a conversation. How can I assist 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. L25N7931-1🕒	0.98	“ geometry”“几何”“ Geometry”
2. L25N5451+1🕒	0.98	“ geometry”“ Geometry”“几何”
3. L27N1036+1🕒	0.98	“几何”“ Geometry”“ geometry”
4. L26N2926+1⛔	0.98	“几何”“ Geometry”“ geometry”
5. L25N3459+1🕒	0.98	“ geometry”“ Geometry”“几何”
6. Broken Compass✅ L30N9187-1	0.98	“ geometry”“ Geometry”“ geomet”
7. L28N6993-1🕒	0.98	“ geometry”“ Geometry”“geometry”
8. L0N5850-1💥	0.97	“ geomet”“ geometric”“<\|im_end\|>”
9. L24N3599-1💤	0.95	“ geometry”“几何”“geometry”
10. L28N3430-1🕒	0.95	“ geomet”“ geometry”“ Geometry”

Shadow specimen, apart from the ten: L20N6408+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 20, neuron 6408, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.68) 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
L20N6408+1🕒 shadow specimen t0-v0.1	1.489	0.49	1.661	1.59	“因为它”“身为”“作为”“ being”“作为中国”“因为其”“ Being”“ as”
L20N6408-1 t0-v0.1	0.653	0.68	0.784	1.14	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.