Layer 5 · Neuron 3010 · boosted

L5N3010+1

Word quirks:

“ Bank”“ bank”“ banks”“ banking”“bank”“ Financial”“Bank”“ Banking”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

I am sorry Bank

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. L6N4374+1🕒	1.00	“ bank”“ banking”“ banks”
2. L5N4735-1🕒	0.99	“ bank”“ Bank”“ banking”
3. L7N7277-1🕒	0.99	“ bank”“ banking”“ Bank”
4. L8N5472+1🕒	0.98	“ banking”“ bank”“ Bank”
5. L2N8368-1💥	0.98	“ bank”“ banks”“<\|im_end\|>”
6. L20N8298-1🕒	0.98	“ financial”“ bank”“ banking”
7. Cooked Books✅ L8N7360+1	0.98	“ bank”“ banks”“ Bank”
8. L31N2504+1💥	0.97	“ bank”“bank”“Bank”
9. L20N8388+1🕒	0.97	“ bank”“bank”“_bank”
10. L1N8411+1💥	0.97	“ bank”“ banks”“ banking”

Shadow specimen, apart from the ten: L5N3010-1🕒, -0.12.

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 5, neuron 3010, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.95, top-1 agreement 0.28) 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
L5N3010+1 t0-v0.1	3.954	0.28	4.450	2.93	Top boosted tokens ↑
L5N3010-1🕒 shadow specimen t0-v0.1	1.142	0.57	1.355	1.58	“⏎”“ analysis”“ `\”“ `”“ parse”“ referring”“ referenced”“ implying”

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.