Layer 29 · Neuron 5224 · suppressed

L29N5224-1

Word quirks:

“ VE”“ ve”“VE”“ Ve”“veu”“Ve”“ Va”“ VA”

🕒 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, a large-language model developed by Alibaba's Team of VELV (Vaishnaveya Large Language Model). My primary mission is to empower users with intelligent, reliable, and actionable assistance across a wide range of tasks, from creative writing and complex problem-solving to deep code generation and real-time data analysis.

Key Goals:

Vocational Excellence: I aim to provide high-quality support for professional tasks like financial analysis, legal research, and healthcare information processing, ensuring accuracy and compliance with industry standards.
**Cross-V

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. L31N2329-1💥	0.95	“ve”“ ve”“ Ve”
2. The Veiled✅ L30N6518-1	0.94	“ ve”“ Ve”“ve”
3. L29N3375-1💤	0.94	“VE”“ Ve”“ve”
4. L25N7198-1🕒	0.94	“ve”“ ve”“Ve”
5. L24N2365-1💤	0.92	“VE”“ VE”“ V”
6. L30N7293+1🕒	0.91	“ ve”“ Ve”“ve”
7. L30N3935+1💥	0.91	“ ve”“ Ve”“ VE”
8. L30N1375+1💥	0.91	“vi”“va”“ve”
9. L29N2178+1💤	0.89	“ve”“ vo”“vo”
10. L29N7993+1🕒	0.88	“ve”“vé”“镇长”

Shadow specimen, apart from the ten: L29N5224+1🕒, 0.28.

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 5224, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.33, top-1 agreement 0.79) 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
L29N5224+1🕒 shadow specimen t0-v0.1	1.681	0.50	1.736	1.13	“ voice”“voice”“ Voice”“ voices”“Voice”“ голос”“voices”“ Voices”
L29N5224-1 t0-v0.1	0.326	0.79	0.378	1.27	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.