Layer 13 · Neuron 3963 · boosted

L13N3963+1

Word quirks:

“ now”“ earlier”“ previously”“ prior”“ NOW”“ before”“ previous”“ later”

🕒 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 mission is to assist users in generating high-quality content, conducting research, and solving complex problems—always prioritizing accuracy and helpfulness. I specialize in multilingual capabilities and can process long contexts when needed. How can I assist you today? Let me know if you'd like to explore my technical strengths further!

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. L12N7605-1💤	0.97	“ previous”“ earlier”“ before”
2. L11N2964-1💤	0.95	“ previously”“ previous”“ now”
3. L11N3666+1💤	0.94	“ now”“ before”“ BEFORE”
4. L12N3600+1💤	0.94	“ now”“ prior”“ before”
5. L14N8272-1🕒	0.94	“ previously”“ prior”“ previous”
6. L9N6157+1🕒	0.94	“ now”“ past”“ previously”
7. L12N2396+1💤	0.94	“ now”“ before”“ previously”
8. L10N499-1🕒	0.94	“ previously”“ prior”“ before”
9. L6N7054+1🕒	0.94	“ before”“ prior”“ previously”
10. L10N4613-1🕒	0.94	“ now”“ previously”“ previous”

Shadow specimen, apart from the ten: L13N3963-1🕒, 0.32.

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 3963, 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
L13N3963+1 t0-v0.1	0.852	0.65	1.074	1.45	Top boosted tokens ↑
L13N3963-1🕒 shadow specimen t0-v0.1	0.983	0.62	1.250	1.35	“ after”“ expected”“ AFTER”“after”“ confirmed”“ follow”“ info”“ clear”

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.