Layer 24 · Neuron 6752 · boosted

L24N6752+1

Word quirks:

“仙”“造的”“控制”“支持的”“控制住”“斑”“aload”“造”

💤 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 Qia, a large-scale scientific research and innovative service model created by the Guangguang Institute of Advanced Information Technology. My mission is to provide high-quality, accurate, and useful information and services across various fields such as science and technology, business, and daily life. I aim to support users in solving complex problems, enhancing decision-making, and exploring new ideas through advanced AI capabilities.

If you have specific questions or tasks, feel free to ask! 🌟

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. L30N4896+1🕒	0.86	“造”“造的”“省”
2. L31N5258-1💥	0.83	“trys”“造”“小人”
3. L30N4497-1💤	0.83	“地上”“喻”“操作中”
4. L29N3971-1🕒	0.83	“能”“可以”“义”
5. L30N5666-1💤	0.82	“开头”“有用”“构成”
6. L25N393+1🕒	0.82	“合”“转”“识别”
7. L29N192-1💤	0.82	“自动”“造”“进一步”
8. L26N5174-1💤	0.82	“置”“增”“观”
9. L24N1539-1💤	0.82	“生成”“表示”“秀”
10. L30N20-1💤	0.82	“搭”“新”“借”

Shadow specimen, apart from the ten: L24N6752-1🕒, 0.39.

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 24, neuron 6752, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.77, top-1 agreement 0.67) 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
L24N6752+1 t0-v0.1	0.773	0.67	0.753	1.24	Top boosted tokens ↑
L24N6752-1🕒 shadow specimen t0-v0.1	1.104	0.64	1.251	1.33	“降”“ Desc”“世”“示”“ mundo”“ desc”“ worlds”“Desc”

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.