Layer 21 · Neuron 4871 · boosted

L21N4871+1

Word quirks:

“ itself”“ themselves”“本身”“自己有”“ yourself”“ oneself”“ herself”“ himself”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qianwen (Wenxin-Yuan) models, a series of advanced AI technologies developed by the Tongsheng (Yun) team of Tongsheng (Yun) Technology. I have no personal identity or mission in the human sense; rather, I am an AI model designed to assist users with tasks like writing, reasoning, and content creation. My core capabilities include:

Advanced Reasoning: Complex problem-solving, scientific calculations, and logical workflows.
Content Creation: Writing, creative writing, and professional documents.
Technical Skills:

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. Off By Ten✅ L23N1533-1	0.99	“ itself”“ themselves”“本身”
2. L15N7229-1🕒	0.98	“ itself”“ themselves”“本身”
3. L18N7416-1🕒	0.97	“ itself”“ herself”“ myself”
4. L20N2310-1🕒	0.97	“ itself”“ themselves”“本身”
5. L20N2905-1🕒	0.97	“ itself”“ themselves”“本身”
6. L17N6458-1💤	0.97	“ itself”“ themselves”“ myself”
7. L17N2549-1💤	0.97	“ itself”“ herself”“ themselves”
8. L19N1447-1🕒	0.97	“ itself”“ themselves”“本身”
9. L18N5976-1💤	0.96	“ itself”“ themselves”“ himself”
10. L19N8894+1💤	0.96	“ itself”“ themselves”“ herself”

Shadow specimen, apart from the ten: L21N4871-1💤, 0.02.

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 21, neuron 4871, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.85, top-1 agreement 0.64) 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
L21N4871+1 t0-v0.1	0.854	0.64	1.019	1.30	Top boosted tokens ↑
L21N4871-1💤 shadow specimen t0-v0.1	0.663	0.71	0.762	1.22	“ pud”“ nicely”“ pur”“值得”“[*”“ managed”“ensee”“ manages”

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.