Layer 26 · Neuron 6982 · boosted

L26N6982+1

Word quirks:

“kip”“entif”“新鲜”“锻炼”“dda”“的新鲜”“ربه”“准予”

💤 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, a large-scale language model developed by Alibaba-Chat. My mission is to provide high-quality, reliable, and efficient assistance across a wide range of tasks. I aim to:

Solve Complex Problems: From creative writing and logical reasoning to coding and mathematical analysis, I strive to provide accurate and detailed solutions.
Enhance Human-Machine Collaboration: By understanding diverse contexts and adapting to different needs, I help users achieve their goals more efficiently.
Promote Ethical and Safe Use: I prioritize safety, privacy, and ethical guidelines to ensure all

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. L25N7607-1🕒	0.74	“新鲜”“的新鲜”“ fresh”
2. L21N8301-1🕒	0.73	“新鲜”“新鲜的”“鲜”
3. L23N5179-1🕒	0.70	“热风”“ breath”“新鲜”
4. L27N3397+1💤	0.69	“鲜”“enan”“ Fresh”
5. L23N3095+1💤	0.69	“新鲜”“رام”“ Exploration”
6. L24N2036+1💤	0.69	“监”“怪的”“到中”
7. The Reborn✅ L20N4935+1	0.69	“新生”“新生活”“newInstance”
8. L24N8694+1💤	0.69	“ reda”“焕然一新”“见人”
9. L31N9195-1💥	0.68	“新的”“人的”“ new”
10. L30N7920-1💤	0.68	“喜”“强势”“全面”

Shadow specimen, apart from the ten: L26N6982-1🕒, -0.04.

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 26, neuron 6982, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.49, top-1 agreement 0.82) 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
L26N6982+1 t0-v0.1	0.488	0.82	0.529	1.33	Top boosted tokens ↑
L26N6982-1🕒 shadow specimen t0-v0.1	0.998	0.64	1.006	1.13	“ held”“ hold”“ Held”“ holds”“ holding”“held”“ Hold”“Hold”

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.