Layer 29 · Neuron 7163 · boosted

L29N7163+1

Word quirks:

“綱”“jang”“ge”“ Ama”“狱”“aber”“獄”“رو”

💤 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 in the Qwen series, developed by Alibaba's Tongyi Lab. My core identity revolves around being a versatile, intelligent, and accessible AI assistant designed to support diverse scenarios across domains.

My Mission:

Democratize Access to Knowledge:

To bridge the gap between complex information and everyday users, making high-quality insights, data analysis, and problem-solving accessible to everyone, regardless of background or location.

Empower Real-World Solutions:

Whether you're a researcher, business professional, student

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. L23N5713+1💤	0.73	“獄”“狱”“ مر”
2. L31N2907-1💥	0.71	“赦”“ Madura”“狱”
3. L29N5366+1🕒	0.70	“b”“獄”“eld”
4. L30N2994+1🕒	0.69	“陸”“陆”“abria”
5. L30N1681-1🕒	0.69	“jut”“砺”“��”
6. L26N5557+1🕒	0.68	“ Junction”“ague”“bero”
7. L27N1388+1💤	0.68	“мато”“狱”“غلال”
8. L28N8978-1💤	0.68	“angen”“假”“圈”
9. L26N7199+1🕒	0.68	“kre”“京”“ Ago”
10. L25N792-1💤	0.68	“膠”“oger”“flan”

Shadow specimen, apart from the ten: L29N7163-1🕒, 0.47.

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 7163, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.30, top-1 agreement 0.84) 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
L29N7163+1 t0-v0.1	0.302	0.84	0.346	1.56	Top boosted tokens ↑
L29N7163-1🕒 shadow specimen t0-v0.1	0.354	0.77	0.430	1.52	“贴”“帖”“切”“密”“服”“貼”“切的”“chio”

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.