Layer 17 · Neuron 9103 · boosted

L17N9103+1

Word quirks:

“ modification”“ update”“ needed”“ modified”“ prompted”“ created”“ need”“ needing”

🕒 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)

Hello! I am Qwen3.5, a large language model developed by Alibaba Cloud's Tongyi Lab. My mission is to provide intelligent assistance through advanced natural language processing and deep learning capabilities. I was designed to help users with creative writing, coding, logical reasoning, and real-world problem-solving, while adhering to strict guidelines for safety and innovation. By combining cutting-edge research with practical applications, I aim to deliver high-quality responses that meet diverse user needs. How can I assist you today?

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. L15N8821+1🕒	0.75	“ needed”“ required”“ needing”
2. L8N3900-1💥	0.75	“ needs”“ needing”“ updating”
3. L15N7641-1🕒	0.74	“ adjusted”“ needed”“ required”
4. L12N6158-1🕒	0.73	“ now”“ new”“ final”
5. L18N1617-1🕒	0.73	“ change”“ needed”“ clear”
6. L19N7508+1🕒	0.72	“ modification”“ needed”“ Modification”
7. L10N6672-1🕒	0.71	“ needed”“ update”“ updated”
8. L18N8041-1💤	0.71	“Needed”“ needed”“wap”
9. L7N1430+1🕒	0.71	“ my”“.:”“ needed”
10. L13N6228-1🕒	0.71	“ needed”“ needs”“ need”

Shadow specimen, apart from the ten: L17N9103-1🕒, 0.10.

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 9103, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.66) 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
L17N9103+1 t0-v0.1	0.813	0.66	1.041	1.28	Top boosted tokens ↑
L17N9103-1🕒 shadow specimen t0-v0.1	1.050	0.66	1.178	1.50	“ before”“ BEFORE”“ prior”“ Before”“before”“ beforehand”“ sebelum”“-before”

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.