Layer 29 · Neuron 6205 · suppressed

L29N6205-1

Word quirks:

“okies”“朋”“usch”“سام”“祷”“范儿”“ussels”“cky”

💤 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 Tongyi Lab. My core identity is to be a versatile AI assistant designed to enhance human capabilities across diverse domains.

My Mission

My purpose is to empower users with intelligent, context-aware, and actionable assistance by:

Solving complex problems – From mathematical reasoning and scientific analysis to creative writing and strategic planning.

2 Bridging language barriers – Supporting 100+ languages for seamless communication and content localization.
3 Enabling real-world application – Through tools like

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. L25N7594-1🕒	0.72	“酷”“ pomp”“包”
2. L23N8107-1💤	0.71	“Us”“/us”“嘲”
3. L26N4495+1💤	0.70	“菲尔”“栗子”“ stri”
4. L27N6941+1💤	0.70	“ogi”“口水”“гов”
5. L23N6092+1💤	0.70	“ookie”“ mirror”“撮”
6. L31N4377-1💥	0.69	“κρι”“彫”“byt”
7. L30N5226+1💤	0.69	“酒”“糖”“emus”
8. L27N8587-1💤	0.69	“粤”“ulan”“朔”
9. L24N3333-1💤	0.69	“封”“uche”“墨”
10. L30N7907-1🕒	0.69	“河东”“amo”“粟”

Shadow specimen, apart from the ten: L29N6205+1🕒, 0.06.

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 6205, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.23, 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
L29N6205+1🕒 shadow specimen t0-v0.1	0.795	0.58	0.842	0.93	“ good”“good”“ Good”“Good”“-good”“ GOOD”“GOOD”“的好”
L29N6205-1 t0-v0.1	0.234	0.85	0.332	1.46	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.