Layer 20 · Neuron 6736 · boosted

L20N6736+1

Word quirks:

“ reasoning”“ reason”“ logic”“ logical”“ judgement”“ razon”“ Reason”“logical”

🕒 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 Q3, a model that I am currently in the process of being developed. My core mission is to provide high-quality, professional, and personalized assistance across various domains, including academic research, logical reasoning, and creative creation. I aim to be an intelligent assistant that can help you solve problems, explore ideas, and make decisions more efficiently.

If you have any specific questions or need help with a particular task, feel free to let me know! We can explore together.

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. L1N5116+1💥	0.91	“<\|im_end\|>”“ reasoning”“ logic”
2. L2N4703+1💥	0.90	“<\|im_end\|>”“ logic”“ logical”
3. L8N3977-1🕒	0.90	“ reasoning”“ logical”“ logic”
4. L19N375+1💤	0.90	“logic”“ logic”“Logic”
5. L2N5271-1💥	0.87	“ reasoning”“<\|im_end\|>”“ reason”
6. L18N2180-1🕒	0.86	“ reasoning”“ reason”“ Reason”
7. L6N8529+1🕒	0.85	“ logic”“ logical”“ logically”
8. L10N6826+1💤	0.85	“ logic”“ logical”“ hypothetical”
9. L13N6567-1🕒	0.84	“ reasoning”“ analysis”“ logic”
10. L23N3060-1🕒	0.84	“ Reason”“ reasoning”“推理”

Shadow specimen, apart from the ten: L20N6736-1🕒, 0.01.

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 20, neuron 6736, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.68, top-1 agreement 0.73) 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
L20N6736+1 t0-v0.1	0.679	0.73	0.689	1.38	Top boosted tokens ↑
L20N6736-1🕒 shadow specimen t0-v0.1	0.640	0.68	0.784	1.31	“ visibility”“ regard”“Std”“olland”“看着”“pert”“甘露”“ visible”

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.