Layer 4 · Neuron 4682 · suppressed

L4N4682-1

Word quirks:

“ laundry”“ clothes”“ washing”“ cleaning”“ wash”“ washed”“ dryer”“ washer”

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

Identity (according to the specimen itself)

I have provided a comprehensive list of items for you.

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. L7N7184+1🕒	0.98	“ dryer”“ laundry”“ washer”
2. L2N4453-1💥	0.92	“<\|im_end\|>”“ washing”“ wash”
3. The Detergent✅ L6N8772-1	0.88	“ detergent”“ soap”“ washing”
4. L28N8732-1🕒	0.87	“ Wash”“ wash”“wash”
5. L25N8546+1🕒	0.87	“ washing”“ washed”“ wash”
6. L26N4400-1🕒	0.86	“wash”“ wash”“ washed”
7. L26N1491-1🕒	0.84	“ wash”“ Wash”“ washing”
8. L6N5844-1🕒	0.83	“ dishwasher”“ washing”“ cleaning”
9. L23N5876+1💤	0.82	“ wash”“ Wash”“ washed”
10. L0N9129+1💥	0.82	“wash”“ wash”“<\|im_end\|>”

Shadow specimen, apart from the ten: L4N4682+1🕒, 0.11.

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 4682, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.46, 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
L4N4682+1🕒 shadow specimen t0-v0.1	1.050	0.61	1.272	1.77	“ ‘”“ protected”“ run”“ management”“ mine”“ referenced”“ resource”“ fish”
L4N4682-1 t0-v0.1	3.461	0.28	3.957	2.49	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.