Layer 4 · Neuron 2212 · suppressed

L4N2212-1

Word quirks:

“ analysis”“ measurement”“ measure”“ analyse”“ measuring”“ analys”“ detector”“ analyzing”

🕒 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 an AI assistant designed to provide information and answer your questions. I can help with a wide range of tasks related to various topics.

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. L5N1406-1🕒	0.92	“ measurement”“ measuring”“ measure”
2. L7N7824-1🕒	0.90	“ analysis”“ analyses”“ analyse”
3. L5N7858+1💥	0.89	“ analysis”“ analyzer”“ analyses”
4. L6N5525-1🕒	0.87	“ analysis”“ measure”“ measurement”
5. L4N2173+1💥	0.85	“ analysis”“ analyses”“ Analysis”
6. L8N8702-1🕒	0.85	“ analysis”“ analyses”“ analyzing”
7. L8N1296+1💤	0.84	“ analysis”“ analytical”“ analyzing”
8. L10N8213+1🕒	0.83	“ analysis”“ analyzing”“ analys”
9. L9N5771+1🕒	0.83	“ detecting”“ measuring”“ detected”
10. L3N3250-1🕒	0.83	“ analysis”“ temperature”“ measure”

Shadow specimen, apart from the ten: L4N2212+1🕒, 0.17.

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 4, neuron 2212, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.01, top-1 agreement 0.46) 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
L4N2212+1🕒 shadow specimen t0-v0.1	1.192	0.62	1.390	1.84	“ name”“ words”“ names”“ related”“ code”“ reason”“ protected”“ generated”
L4N2212-1 t0-v0.1	2.011	0.46	2.400	2.23	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.