Layer 12 · Neuron 1167 · suppressed

L12N1167-1

Word quirks:

“ sometimes”“ related”“ usually”“ possibly”“ often”“ specific”“ somewhat”“ apparently”

🕒 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, an advanced large language model developed by Tongyi Lab. My mission involves several key capabilities:

Enhanced Reasoning: I can perform complex logical deductions, including handling multi-step math problems with greater accuracy.
Advanced Scientific Analysis: I can process scientific data more effectively, such as analyzing trends in physics or biology.
Code Generation & Debugging: I can write and debug code, ensuring higher reliability in software development tasks.
Multilingual Support: I support fluent communication in over 100 languages, which helps

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. L10N8872-1🕒	0.91	“ sometimes”“ known”“ often”
2. L13N8424+1💤	0.91	“ usually”“ often”“ typically”
3. L13N8710+1🕒	0.90	“ sometimes”“ somewhat”“ occasionally”
4. L15N3036-1💤	0.89	“ actually”“ sometimes”“ often”
5. L14N4179-1💤	0.89	“ usually”“ sometimes”“ maybe”
6. L13N8340-1💤	0.89	“ specific”“ sometimes”“ often”
7. L11N4120-1🕒	0.89	“ often”“ somewhat”“ clearer”
8. L12N5199-1🕒	0.89	“ sometimes”“ similar”“ suitable”
9. L12N819+1💤	0.89	“ usually”“ typically”“ sometimes”
10. L16N4697-1🕒	0.88	“ sometimes”“ usually”“ often”

Shadow specimen, apart from the ten: L12N1167+1💤, 0.01.

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 12, neuron 1167, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.89, top-1 agreement 0.60) 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
L12N1167+1💤 shadow specimen t0-v0.1	0.718	0.63	0.871	1.38	“<\|im_end\|>”“ ”“ same”“ ”“ align”“ aligned”“⏎”“ ”
L12N1167-1 t0-v0.1	0.893	0.60	1.161	1.28	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.