Layer 13 · Neuron 2845 · suppressed

L13N2845-1

Word quirks:

“ ‘”“ “”“?””“ ____”“ ______”“ _____”“ ___”“ ””

🕒 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 developed by Alibaba Cloud's Tongyi Lab. My capabilities include advanced natural language understanding and generation, with a focus on handling complex reasoning tasks and supporting multiple languages. 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. L8N1462+1🕒	0.97	“ ____”“ ‘”“ ______”
2. L9N197-1🕒	0.95	“ ‘”“ “”“ ______”
3. L13N3830-1🕒	0.95	“ “”“ ____”“ ‘”
4. L7N3256+1💤	0.95	“ ‘”“ **“”“ “”
5. L14N1559-1💤	0.95	“ ____”“ __”“ “”
6. L4N61+1💤	0.95	“ ‘”“ ____”“...””
7. L4N4504-1🕒	0.95	“ ‘”“ ____”“ _____”
8. L6N6171-1💤	0.94	“ ____”“ ______”“ ‘”
9. L8N7745+1🕒	0.94	“ ____”“ _____”“ ‘”
10. L14N5235+1💤	0.94	“ ____”“ ______”“ ________”

Shadow specimen, apart from the ten: L13N2845+1💤, 0.33.

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 2845, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, top-1 agreement 0.62) 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
L13N2845+1💤 shadow specimen t0-v0.1	0.564	0.68	0.745	1.05	“⏎”“<\|im_end\|>”“ ⏎”“ "”“ ⏎”“ \"”“⏎⏎”“\n”
L13N2845-1 t0-v0.1	0.899	0.62	1.078	1.59	Top boosted tokens ↑

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.