Layer 29 · Neuron 3375 · suppressed

L29N3375-1

Word quirks:

“VE”“ Ve”“ve”“veil”“Ve”“ Veg”“ ve”“VG”

💤 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 Qwen3.5, a large language model developed by Tongyi Lab. My primary mission is to empower users with reliable, efficient, and intelligent AI assistance across a wide range of tasks. Here's how I aim to deliver on this mission:

Core Capabilities & Mission Focus:

Verges of Human-Level Reasoning

I excel in complex reasoning (math, science, logic) and can break down multi-step problems into clear, logical steps—making me a strong companion for academic research, technical problem-solving, and decision-making.

**Deep

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. L25N7198-1🕒	0.97	“ve”“ ve”“Ve”
2. The Veiled✅ L30N6518-1	0.96	“ ve”“ Ve”“ve”
3. L31N2329-1💥	0.95	“ve”“ ve”“ Ve”
4. L29N5224-1🕒	0.94	“ VE”“ ve”“VE”
5. L24N2365-1💤	0.93	“VE”“ VE”“ V”
6. L30N7293+1🕒	0.91	“ ve”“ Ve”“ve”
7. L30N5163-1🕒	0.90	“ Ve”“ve”“ ve”
8. L29N2178+1💤	0.88	“ve”“ vo”“vo”
9. L30N3935+1💥	0.87	“ ve”“ Ve”“ VE”
10. L30N1375+1💥	0.86	“vi”“va”“ve”

Shadow specimen, apart from the ten: L29N3375+1🕒, 0.36.

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 3375, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.24, top-1 agreement 0.81) 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
L29N3375+1🕒 shadow specimen t0-v0.1	0.337	0.81	0.372	1.24	“ vehicles”“ vehicle”“ Vehicle”“vehicle”“ Vehicles”“vehicles”“車の”“Vehicle”
L29N3375-1 t0-v0.1	0.244	0.81	0.304	1.30	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.