Layer 29 · Neuron 264 · suppressed

L29N264-1

Word quirks:

“цов”“削”“紫”“炸”“ cub”“ômes”“oso”“ÁL”

💤 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, the latest large language model developed by Alibaba's Tongyi Lab. My core purpose is to empower users with advanced capabilities across a wide range of tasks, from reasoning and coding to visual analysis and multi-language support.

My Mission

My mission is to bridge the gap between human intent and AI execution by:

Solving complex problems: Whether it's debugging code, analyzing scientific data, or planning intricate workflows, I aim to provide accurate, logical, and actionable solutions.
Enhancing human potential: I design interactions that are **natural,

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. L29N4884+1🕒	0.76	“ál”“犁”“erso”
2. L30N8295+1💤	0.73	“coil”“弹”“�”
3. L29N6965+1💤	0.73	“ Levin”“狞”“ogg”
4. L27N7558-1💤	0.72	“钉”“九”“筑”
5. L28N3728+1💤	0.72	“鸭”“猎”“utum”
6. L23N9125-1💤	0.72	“斗”“باد”“贡”
7. L30N3880+1💤	0.71	“онер”“uge”“减”
8. L29N1995+1💤	0.71	“znak”“ steep”“等相关”
9. L24N5904+1💤	0.71	“猫”“ Ox”“貓”
10. L26N769-1💤	0.71	“宗”“inser”“红”

Shadow specimen, apart from the ten: L29N264+1💥, 0.19.

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 264, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.22, top-1 agreement 0.85) 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
L29N264+1💥 shadow specimen t0-v0.1	10.810	0.00	10.895	0.56	“集体”“ collective”“ collectively”“ Collective”“collect”“集體”“集体的”“ коллек”
L29N264-1 t0-v0.1	0.220	0.85	0.238	1.40	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.