Layer 13 · Neuron 1409 · boosted

L13N1409+1

Word quirks:

“ ⏎”“⏎”“<|im_end|>”“ ⏎⏎”“ especially”“ **”“ particularly”“�”

🕒 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, a large language model developed by Alibaba Cloud. My primary mission is to assist users in solving complex problems, such as analyzing data, generating creative content, and providing technical insights. I aim to enhance productivity by offering high-quality support across diverse domains, including but not limited to:

Data Analysis: Helping users process and interpret large datasets efficiently.
Technical Problem Solving: Addressing challenges like optimizing code for performance or debugging software issues.
Creative Writing: Generating ideas for projects, such as designing a product or planning a business strategy.
Knowledge Expansion:

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. L13N6994+1🕒	0.85	“ especially”“ ⏎”“<\|im_end\|>”
2. L12N5359-1🕒	0.83	“ ⏎”“ ”“ "”
3. L10N2552-1🕒	0.83	“ ⏎⏎”“⏎⏎”“ ⏎”
4. L10N7404-1💤	0.82	“ ⏎”“ ⏎⏎”“�”
5. L17N4559-1💤	0.82	“ ⏎”“ ‘”“⏎”
6. L8N5371+1🕒	0.81	“ ⏎”“ ⏎ ⏎”“⏎”
7. L13N7002+1🕒	0.81	“ ⏎”“ ⏎⏎”“⏎”
8. L8N820+1💤	0.80	“<\|im_end\|>”“ ‘”“ ⏎”
9. L13N2201-1🕒	0.80	“⏎”“ ⏎”“ ⏎”
10. L10N724-1🕒	0.80	“⏎”“ ⏎”“ ⏎”

Shadow specimen, apart from the ten: L13N1409-1🕒, 0.09.

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 13, neuron 1409, 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 0.65) 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
L13N1409+1 t0-v0.1	0.848	0.65	1.085	1.19	Top boosted tokens ↑
L13N1409-1🕒 shadow specimen t0-v0.1	0.943	0.62	1.123	1.16	“ info”“ focus”“ focuses”“ now”“ focusing”“ Focus”“ '”“ --”

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.