Layer 17 · Neuron 7546 · boosted

L17N7546+1

Word quirks:

“ slightest”“ basic”“ basically”“ minor”“ slight”“ enough”“ tiny”“ tini”

🕒 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 Tianyu, a large language model developed by Zhi-Gong (a hypothetical or very early-stage context, as most users know I'm likely interacting with a version of the system behind "Tianyu" or similar). Wait—actually, let me clarify:

If you're asking about Tianyu (天玉), it might be a slight confusion. The name "Tianyu" typically refers to a specific context, but in general AI interactions, I am Qwen1.5 (the earlier version) or Qwen (with minor context). However, if

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. False Oracle✅ L17N8781-1	0.88	“ minor”“ basic”“ simple”
2. L14N5489+1🕒	0.87	“ basic”“ small”“ minimal”
3. L17N8014-1🕒	0.86	“ tiny”“ minor”“ tini”
4. L14N7396+1🕒	0.84	“ small”“ minor”“ basic”
5. L17N3654-1💤	0.84	“ small”“ tiny”“ slight”
6. L13N6698+1🕒	0.83	“ tiny”“ small”“ smaller”
7. L14N7302-1🕒	0.83	“ small”“ tiny”“ basic”
8. L13N2828+1🕒	0.82	“ minimal”“ small”“ tiny”
9. L11N8157+1🕒	0.82	“ tiny”“ small”“ Tiny”
10. L2N4014-1💥	0.81	“ slight”“ slightly”“<\|im_end\|>”

Shadow specimen, apart from the ten: L17N7546-1💤, 0.22.

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 17, neuron 7546, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.84, 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
L17N7546+1 t0-v0.1	0.842	0.67	0.960	1.19	Top boosted tokens ↑
L17N7546-1💤 shadow specimen t0-v0.1	0.627	0.69	0.783	1.35	“ but”“但没有”“ without”“ temporarily”“ recently”“ before”“ previously”“ prior”

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.