Layer 16 · Neuron 2515 · boosted

L16N2515+1

Word quirks:

“ about”“ regarding”“ stating”“ saying”“ seeming”“ concerning”“ specifically”“ specifying”

🕒 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 Qwen3.5, the latest large language model in the T702 series. My primary goal is to assist users like you to provide with high-quality capabilities. I aim to help users to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high-quality capabilities. I aim to provide with high

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. L7N5933-1🕒	0.75	“ my”“ saying”“ stating”
2. L12N4628+1💤	0.74	“ specific”“ specifically”“ implying”
3. L20N8582-1🕒	0.73	“ regarding”“ concerning”“ managing”
4. L6N8420+1💥	0.72	“<\|im_end\|>”“ concerning”“ about”
5. L16N4029-1🕒	0.72	“ specifics”“ regarding”“ regards”
6. L16N9071+1🕒	0.72	“ telling”“ knowing”“ saying”
7. L13N1778+1🕒	0.72	“ specific”“ specifics”“ specifies”
8. L14N953+1💤	0.72	“ specified”“ description”“ indicating”
9. L13N5587+1🕒	0.71	“ specific”“ saying”“ supposedly”
10. L10N6392-1🕒	0.71	“ basically”“ emphas”“ describing”

Shadow specimen, apart from the ten: L16N2515-1💤, 0.29.

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 2515, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.77, top-1 agreement 0.63) 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
L16N2515+1 t0-v0.1	0.767	0.63	0.956	1.43	Top boosted tokens ↑
L16N2515-1💤 shadow specimen t0-v0.1	0.464	0.72	0.628	1.38	“ thus”“ therefore”“ however”“ meanwhile”“ this”“ here”“?\”“ makes”

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.