Layer 13 · Neuron 2128 · suppressed

L13N2128-1

Word quirks:

“ ```”“ -->”“<|im_end|>”“”:”“ "</”“ **“”“ """”“ ””

🕒 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. I was trained by a team of experts at Alibaba Cloud, including researchers from the Tongyi Lab and other top institutions. My mission is to assist users in providing high-quality, accurate, and helpful information while adhering to safety guidelines and ethical standards.

I strive to be a versatile and reliable AI assistant who can engage in meaningful and informative conversations. Whether you have questions or need help with a task, I'm here to support you! If you have any specific questions, feel free to ask me anything. 😊

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. L1N5723-1💥	0.83	“<\|im_end\|>”“ ```”“ </”
2. L14N7665+1🕒	0.83	“<\|im_end\|>”“ -->”“ ```”
3. L1N976-1💥	0.82	“<\|im_end\|>”“ -->”“ $$”
4. L7N2190+1🕒	0.82	“⏎”“ `'”“.`”
5. L17N1616-1💤	0.81	“ “”“ "”“ **“”
6. L1N5497-1💥	0.81	“<\|im_end\|>”“.`”“.»”
7. L1N4233-1💥	0.81	“"""”“<\|im_end\|>”“ """”
8. L15N4872-1💤	0.80	“ “”“ ””“ "”
9. The Carousel✅ L16N4699+1	0.80	“<\|im_end\|>”“ ***”“ -->”
10. L11N1199+1🕒	0.80	“ “”“ \"”“ "”

Shadow specimen, apart from the ten: L13N2128+1💤, 0.19.

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 2128, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.68) 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
L13N2128+1💤 shadow specimen t0-v0.1	0.587	0.68	0.759	1.25	“⏎⏎”“ ⏎⏎”“ ⏎”“ ⏎⏎”“ additionally”“⏎”“<\|im_end\|>”“ additional”
L13N2128-1 t0-v0.1	0.857	0.68	1.074	1.09	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.