Layer 16 · Neuron 2186 · suppressed

L16N2186-1

Word quirks:

“ compared”“ comparable”“ your”“ similar”“ equivalent”“ compare”“ Compared”“ yours”

🕒 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, your AI assistant developed by me. My mission is to provide high-quality support across a wide range of capabilities—helping users with accurate information, deep analysis, and problem-solving. How can I assist you? 😊

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. L17N5562+1🕒	0.90	“ compared”“ similar”“ comparable”
2. L14N183-1🕒	0.90	“ similar”“ same”“ similarly”
3. L17N8796+1🕒	0.89	“ similar”“ comparable”“ compared”
4. L17N8948+1💤	0.88	“ comparable”“ similar”“ analogous”
5. L10N4742+1🕒	0.88	“ comparison”“ comparable”“ comparing”
6. L17N7820+1💤	0.88	“ comparison”“ comparable”“ compared”
7. L15N7693-1🕒	0.88	“ compared”“ similar”“ than”
8. L14N5767-1🕒	0.88	“ compared”“ comparable”“ unlike”
9. L9N7516+1🕒	0.87	“ comparison”“ comparisons”“ similar”
10. L12N4778+1🕒	0.86	“ comparable”“ similar”“ comparison”

Shadow specimen, apart from the ten: L16N2186+1💤, 0.17.

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 16, neuron 2186, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.87, top-1 agreement 0.59) 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
L16N2186+1💤 shadow specimen t0-v0.1	0.506	0.75	0.677	1.20	“ means”“ ALWAYS”“ ONLY”“ always”“ focus”“ focuses”“ they”“ yes”
L16N2186-1 t0-v0.1	0.865	0.59	1.074	1.26	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.