Layer 30 · Neuron 4012 · boosted

L30N4012+1

Word quirks:

“彦”“昌”“玮”“璋”“骏”“晖”“葆”“sker”

🕒 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)

Hi, I am Qwen3.5, a large language model developed by Tong Y ali Cloud. My primary mission is to empower users with intelligent, accurate, and meaningful assistance across a wide range of tasks. Whether you need help with complex reasoning, creative writing, data analysis, or real-world problem-solving, I aim to provide clear, context-aware responses that adapt to your needs. How can I assist you today? 😊

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. L31N5934-1💥	0.89	“瑞”“虏”“宇”
2. L30N8450+1💤	0.88	“皓”“朱”“广”
3. L31N4324+1💥	0.88	“尤”“浩”“富”
4. L28N3671+1💤	0.88	“希”“御”“沙”
5. L30N5537-1💤	0.87	“景”“凌”“顾”
6. L27N8150-1💤	0.87	“眠”“=~”“彰”
7. L28N7891-1💤	0.87	“品”“佩”“志”
8. L29N3579-1🕒	0.86	“贺”“永定”“坐”
9. L28N6339+1💤	0.86	“太”“芮”“玉”
10. L28N3288-1🕒	0.86	“维”“阳”“象”

Shadow specimen, apart from the ten: L30N4012-1🕒, -0.08.

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 30, neuron 4012, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.46, 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
L30N4012+1 t0-v0.1	0.463	0.79	0.525	2.24	Top boosted tokens ↑
L30N4012-1🕒 shadow specimen t0-v0.1	0.441	0.75	0.521	1.67	“out”“ie”“ive”“ies”“top”“list”“let”“ance”

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.