Layer 28 · Neuron 2313 · suppressed

L28N2313-1

Word quirks:

“ ”“ XX”“ XXX”“X”“5”“ xx”“ fifty”“ X”

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

Identity (according to the specimen itself)

The reply didn't hold together

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1

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. L26N6285-1🕒	0.88	“ XX”“XX”“ XXX”
2. L0N6065+1💥	0.86	“<\|im_end\|>”“ xx”“ xxx”
3. L0N8082+1💥	0.86	“<\|im_end\|>”“ xx”“x”
4. L27N946-1🕒	0.85	“-X”“ XX”“-${”
5. L25N435+1🕒	0.85	“ XX”“ XV”“XX”
6. L0N9041-1💥	0.85	“<\|im_end\|>”“ xx”“ XX”
7. L0N4470-1💥	0.84	“ x”“<\|im_end\|>”“ xa”
8. L28N2818-1🕒	0.84	“.x”“xx”“.X”
9. L0N6068-1💥	0.83	“ x”“ X”“ xx”
10. L0N6543-1💥	0.83	“_x”“ x”“X”

Shadow specimen, apart from the ten: L28N2313+1💤, 0.22.

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 28, neuron 2313, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.09, top-1 agreement 0.19) 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
L28N2313+1💤 shadow specimen t0-v0.1	0.316	0.84	0.333	1.23	“ \(”“ $”“\(”“-$”“ <$”“科”“$”“ \"$”
L28N2313-1 t0-v0.1	2.085	0.19	2.143	0.86	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.