Layer 12 · Neuron 5071 · suppressed

L12N5071-1

Word quirks:

“ ____”“ ***”“<|im_end|>”“⏎”“⏎⏎”“ ****”“ ______”“ _____”

🕒 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 a large language model trained by Google. My mission is to provide high-quality, accurate, and helpful information to users. I strive to ensure that my responses are clear, concise, and easy to understand.

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. L11N2781-1🕒	1.00	“⏎”“⏎⏎”“<\|im_end\|>”
2. L10N382+1🕒	0.98	“<\|im_end\|>”“ ***”“ ____”
3. L10N4796+1💤	0.98	“⏎”“<\|im_end\|>”“ ***”
4. L8N5059+1🕒	0.98	“<\|im_end\|>”“⏎”“ ____”
5. L9N790+1🕒	0.97	“⏎”“<\|im_end\|>”“ _____”
6. L6N741-1💤	0.97	“⏎”“<\|im_end\|>”“ ***”
7. L9N513+1🕒	0.96	“<\|im_end\|>”“ ____”“ ______”
8. The Stencil✅ L6N2135-1	0.96	“<\|im_end\|>”“ ***”“ _____”
9. L10N4754-1💤	0.96	“ ***”“ ____”“ ______”
10. L13N5988+1🕒	0.96	“⏎”“ ***”“ ____”

Shadow specimen, apart from the ten: L12N5071+1🕒, 0.34.

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 12, neuron 5071, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.92, top-1 agreement 0.64) 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
L12N5071+1🕒 shadow specimen t0-v0.1	0.834	0.64	1.064	1.16	“ ‘”“�”“ https”“ ⏎”“ ⏎⏎”“ --”“ #”“ ~”
L12N5071-1 t0-v0.1	0.924	0.64	1.111	1.48	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.