Layer 1 · Neuron 5866 · boosted

L1N5866+1

Word quirks:

“ reload”“ refresh”“<|im_end|>”“ update”“ refreshed”“ refreshing”“ reloading”“ Refresh”

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

Identity (according to the specimen itself)

I'm ready

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. L22N5861+1🕒	0.89	“ refreshed”“ refresh”“ renewal”
2. L3N8189+1💥	0.89	“ refresh”“ renewal”“<\|im_end\|>”
3. L22N6712-1🕒	0.88	“ refresh”“-refresh”“ refreshing”
4. L19N3050+1💤	0.87	“ refreshing”“-refresh”“ refresh”
5. L6N5145-1🕒	0.87	“ refreshed”“ refresh”“ update”
6. L20N6296-1🕒	0.86	“ refresh”“refresh”“ Refresh”
7. L19N6881+1🕒	0.82	“ fresh”“重新开始”“fresh”
8. L21N8735-1🕒	0.81	“ renewal”“ renew”“ Renew”
9. L18N6384-1🕒	0.81	“ restart”“ refresh”“ restarting”
10. L6N6156-1🕒	0.80	“ redesign”“ refreshed”“ update”

Shadow specimen, apart from the ten: L1N5866-1💥, 0.01.

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 1, neuron 5866, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 5.22, top-1 agreement 0.15) 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
L1N5866+1 t0-v0.1	5.219	0.15	5.753	2.86	Top boosted tokens ↑
L1N5866-1💥 shadow specimen t0-v0.1	3.204	0.24	3.621	2.98	“<\|im_end\|>”“ python”“⏎”“⏎⏎”“⏎ ⏎”“python”“.`”“'?”

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.