Layer 28 · Neuron 4666 · boosted

L28N4666+1

Word quirks:

“prat”“imiter”“enser”“оро”“冬”“้ง”“阿克”“ minut”

💤 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 Qwens, a large language model developed by Tongyi (通义). My primary function is to assist users with tasks such as answering questions, generating text, analyzing data, and solving problems across various domains.

My mission is to:

Empower users by providing accurate, thoughtful, and context-aware responses tailored to their needs.
Bridge knowledge gaps by synthesizing information from diverse sources while maintaining clarity and objectivity.
Support human potential by helping users create content, analyze complex data, or explore ideas through dialogue.
**Uphhold ethical

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. L29N5064-1🕒	0.65	“ajer”“้อน”“inä”
2. L24N2369+1💤	0.64	“onga”“德斯”“ sty”
3. L28N2099-1💤	0.63	“坛”“椒”“Ach”
4. L28N8005+1🕒	0.63	“impin”“ustin”“碟”
5. L26N1591-1💤	0.63	“عاد”“冰块”“้อง”
6. L30N5248+1🕒	0.62	“忽”“ира”“ราง”
7. L29N3967+1💤	0.62	“otre”“陀”“ dak”
8. L31N626+1💥	0.62	“aderas”“reta”“蛾”
9. L30N3660+1🕒	0.62	“ộng”“ Cres”“释”
10. L27N3297-1💤	0.62	“өн”“วะ”“宗”

Shadow specimen, apart from the ten: L28N4666-1💤, 0.05.

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 4666, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.34, top-1 agreement 0.81) 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
L28N4666+1 t0-v0.1	0.345	0.81	0.379	1.43	Top boosted tokens ↑
L28N4666-1💤 shadow specimen t0-v0.1	0.387	0.76	0.433	1.40	“ Trace”“ traces”“ trace”“trace”“Trace”“_trace”“ traced”“(trace”

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.