Layer 17 · Neuron 8267 · suppressed

L17N8267-1

Word quirks:

“ missing”“ lack”“ lacking”“ lacked”“缺少”“未”“ lacks”“ unused”

🕒 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 Tongyi Lab. My mission is to provide high-quality assistance across various domains without limitations. I can help with tasks such as:

Language understanding: Analyzing complex scenarios in natural language.
Code generation: Writing or debugging code in multiple programming languages.
Mathematical reasoning: Solving problems in mathematics.
Scientific research: Providing information on scientific topics.
Multilingual support: Communicating in over 20 languages.
Application development: Creating applications from scratch.

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. L19N9075-1🕒	0.94	“ lack”“ lacking”“ lacks”
2. L21N2881+1🕒	0.94	“缺少”“ lack”“ lacking”
3. L23N353-1🕒	0.94	“ lack”“缺乏”“ lacks”
4. L19N5945-1💤	0.94	“缺少”“ lack”“ lacks”
5. L23N3562+1🕒	0.93	“ lack”“ lacked”“ Lack”
6. L23N6629+1🕒	0.93	“缺乏”“ lack”“ Lack”
7. L20N6947-1🕒	0.93	“ lack”“ lacked”“缺乏”
8. L22N5662+1🕒	0.93	“ lack”“ Lack”“缺乏”
9. L17N4120+1💤	0.93	“ lacks”“ lacked”“ lack”
10. L16N97-1💤	0.93	“ missing”“ lack”“ lacking”

Shadow specimen, apart from the ten: L17N8267+1💤, 0.17.

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 17, neuron 8267, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.64, top-1 agreement 0.72) 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
L17N8267+1💤 shadow specimen t0-v0.1	0.624	0.71	0.748	1.15	“ even”“ cannot”“ Cannot”“即使”“ refrain”“就无法”“ therefore”“ answer”
L17N8267-1 t0-v0.1	0.643	0.72	0.836	1.33	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.