Layer 26 · Neuron 6285 · suppressed

L26N6285-1

Word quirks:

“ XX”“XX”“ XXX”“XXX”“XXXX”“X”“ LX”“ X”

🕒 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 XIII, a large language model developed by XXIX. My mission is to assist users in solving complex problems, engaging in creative thinking, and providing accurate information across a wide range of domains. Whether you need help with coding, writing, analysis, or just a thought-provoking conversation, I aim to provide clear, thoughtful, and helpful responses while respecting ethical guidelines and ensuring safety. How can I assist you today?

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. L30N3534-1💥	0.93	“xx”“xxx”“XX”
2. L0N6065+1💥	0.90	“<\|im_end\|>”“ xx”“ xxx”
3. L28N2818-1🕒	0.90	“.x”“xx”“.X”
4. L0N8082+1💥	0.89	“<\|im_end\|>”“ xx”“x”
5. L0N9105+1💥	0.88	“x”“X”“xxx”
6. L0N8411-1💥	0.88	“xx”“<\|im_end\|>”“xxx”
7. L0N8744+1💥	0.88	“<\|im_end\|>”“ xxx”“x”
8. L28N2313-1💥	0.88	“ ”“ XX”“ XXX”
9. L0N9041-1💥	0.87	“<\|im_end\|>”“ xx”“ XX”
10. L24N5540+1💤	0.86	“ XX”“XC”“XX”

Shadow specimen, apart from the ten: L26N6285+1💤, 0.07.

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 26, neuron 6285, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.71, top-1 agreement 0.76) 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
L26N6285+1💤 shadow specimen t0-v0.1	0.458	0.76	0.500	1.41	“署”“嚏”“ presto”“untas”“窟”“ darm”“Sn”“隧道”
L26N6285-1 t0-v0.1	0.713	0.76	0.693	1.18	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.