Layer 24 · Neuron 1454 · suppressed

L24N1454-1

Word quirks:

“ rising”“ rises”“ rise”“ Rise”“ Rising”“rising”“上升”“ rose”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Gövy, a large-scale AI model trained by the Gövy Team. My mission is to provide accurate, efficient, and helpful assistance across a wide range of tasks—whether it's answering questions, creating content, analyzing data, or solving complex problems. I aim to grow steadily in terms of comprehensive, precise, and creative capabilities while maintaining a strong foundation in ethical and safe interactions.

If you have specific tasks or need help with something, feel free to ask! 🚀

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. L31N3189-1💥	0.99	“ rise”“ rose”“ rising”
2. L23N3226+1🕒	0.98	“ rise”“ Rise”“-rise”
3. L25N5400+1🕒	0.97	“ rise”“ Rise”“上升”
4. L24N5535+1🕒	0.97	“ rise”“ Rise”“ rises”
5. L23N5020+1🕒	0.96	“ rising”“上升”“ rise”
6. L0N6536-1💥	0.94	“<\|im_end\|>”“ Rise”“ rise”
7. L24N2626+1🕒	0.94	“上升”“ rise”“ Rise”
8. L24N8553+1🕒	0.94	“上升”“ rise”“ rising”
9. L27N945-1🕒	0.93	“ rise”“ rises”“ Rise”
10. L22N3794-1💥	0.93	“ rise”“ Rise”“rise”

Shadow specimen, apart from the ten: L24N1454+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 24, neuron 1454, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.56, top-1 agreement 0.72) 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
L24N1454+1🕒 shadow specimen t0-v0.1	1.512	0.61	1.555	1.38	“改善”“Impro”“ improvement”“ deep”“ Improvement”“ improving”“深”“ improves”
L24N1454-1 t0-v0.1	0.558	0.72	0.648	1.17	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.