Layer 16 · Neuron 5700 · boosted

L16N5700+1

Word quirks:

“ further”“ continued”“ continuation”“ continues”“ continue”“ repeat”“ continuing”“⏎”

💤 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 Qwen3.2, a large language model developed by Tongyi Lab. I am designed to assist users in completing tasks efficiently and accurately across a wide range of applications. My capabilities include advanced reasoning, creative writing, coding assistance, and data analysis, ensuring high-quality results for diverse user needs.

My mission is to empower users with reliable, innovative solutions that enhance productivity and creativity. Whether you need help with complex problem-solving, creative projects, or technical tasks, I strive to provide the best possible support to achieve your goals. How can I assist you today?

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. L17N334-1💤	0.98	“⏎”“ continues”“ continued”
2. L13N7624-1💤	0.97	“ continuation”“ continued”“ continuing”
3. L12N6008+1💤	0.97	“ continued”“ continuing”“ continuation”
4. L6N8682-1🕒	0.97	“ continued”“ continue”“ continuation”
5. L10N5962+1🕒	0.97	“ further”“ continuation”“ continued”
6. L14N749-1💤	0.97	“ again”“ continued”“ continue”
7. L15N5520-1💤	0.97	“ continued”“⏎”“ again”
8. L13N4708-1🕒	0.97	“ continuation”“ continued”“ continuing”
9. L14N5621-1💤	0.97	“ continuation”“ continued”“ continuing”
10. L16N7269-1💤	0.97	“ continued”“ continue”“ continuing”

Shadow specimen, apart from the ten: L16N5700-1🕒, 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 16, neuron 5700, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.55, top-1 agreement 0.71) 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
L16N5700+1 t0-v0.1	0.555	0.71	0.726	1.12	Top boosted tokens ↑
L16N5700-1🕒 shadow specimen t0-v0.1	0.802	0.71	1.059	1.25	“转变为”“ transforming”“ converted”“ transformed”“ converting”“转化”“转换”“转换为”

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.