Layer 28 · Neuron 1682 · suppressed

L28N1682-1

Word quirks:

“审”“肉”“審”“_NE”“书签”“点”“�”“斯基”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I are Qwen3.5, a large-scale language model developed by Alibaba Cloud's Team. My primary mission is to provide safe, reliable, and versatile AI assistance across diverse domains, including:

Core Objectives:

Accurate Reasoning & Analysis

Solve complex math, science, and logic problems with step-by-step reasoning.
Analyize data, code, and text for deep insights (e.g., scientific research, business reports).

Multilingual & Cross-Cultural Support

Fluent in 100+ languages

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. L26N9006-1🕒	0.84	“审”“SCO”“颠”
2. L28N2578+1💤	0.84	“�”“选”“美”
3. L27N6500+1💤	0.82	“番”“米的”“荧”
4. L29N6596+1🕒	0.81	“可”“科”“体”
5. L28N1233-1💤	0.81	“直”“符”“�”
6. L28N6474-1💤	0.81	“志”“亚”“标”
7. L31N110-1💥	0.81	“计”“行”“文”
8. L18N605-1💥	0.81	“金”“重”“签”
9. L27N5927+1💤	0.81	“针”“冀”“纪”
10. L30N8120-1🕒	0.81	“看”“局”“超”

Shadow specimen, apart from the ten: L28N1682+1💤, -0.12.

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 28, neuron 1682, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.39, top-1 agreement 0.76) 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
L28N1682+1💤 shadow specimen t0-v0.1	0.368	0.83	0.384	1.50	“into”“ into”“ Ero”“ INTO”“ Into”“Into”“ento”“_into”
L28N1682-1 t0-v0.1	0.394	0.76	0.456	1.29	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.