Layer 29 · Neuron 4928 · suppressed

L29N4928-1

Word quirks:

“ N”“李”“ ن”“ n”“ern”“ н”“ niz”“能”

🕒 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 Qwen3.5, the latest NLP model in the Nernu series. My mission is to empower users with Nernu's advanced capabilities by providing Nernu's high-quality, Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu's Nernu

The AI judge wasn't sure this holds together.

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. L28N8519-1🕒	0.91	“ n”“李”“ н”
2. Stuck N Key✅ L27N4405+1	0.90	“ н”“ $”“ N”
3. L26N7797-1🕒	0.89	“ N”“李”“ n”
4. L29N6582-1🕒	0.88	“ n”“.n”“ н”
5. L30N3874+1🕒	0.88	“ n”“ н”“{n”
6. L26N7430+1🕒	0.88	“ N”“N”“_N”
7. L26N8189+1🕒	0.88	“ N”“N”“-N”
8. L31N1144+1💥	0.87	“ N”“N”“李”
9. L29N7190+1🕒	0.87	“ ن”“李”“>N”
10. L26N7033-1🕒	0.87	“ N”“*N”“Ｎ”

Shadow specimen, apart from the ten: L29N4928+1💤, 0.51.

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 29, neuron 4928, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.57, top-1 agreement 0.69) 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
L29N4928+1💤 shadow specimen t0-v0.1	0.167	0.88	0.198	1.19	“(nav”“(ns”“(nr”“(nx”“/ns”“ na”“/ne”“ нар”
L29N4928-1 t0-v0.1	0.565	0.69	0.647	1.08	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.