Layer 16 · Neuron 9183 · suppressed

L16N9183-1

Word quirks:

“ specific”“ differently”“ specifics”“ different”“ three”“ Specific”“ specifically”“ distinct”

🕒 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 with a specific purpose: to understand, guide, and assist in distinct ways. Here's what that means for me:

Who Am I?

Identity: I am not a physical entity but a text-based model. My existence is defined by my ability to communicate.
Capabilities:
Language Understanding: I can process and respond in 108 languages, from code to symbols.
Task-Specific Functions:
Math: I can solve problems in a specific domain (e.g

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. L10N6053+1🕒	0.94	“ specific”“ particular”“ different”
2. L13N6266-1💤	0.91	“ specific”“ different”“ differing”
3. L12N3574-1🕒	0.91	“ specific”“ different”“ Specific”
4. L12N4431+1💤	0.90	“ specific”“ specifics”“ specifically”
5. L16N6487-1💤	0.89	“ specifics”“ specific”“ different”
6. L21N545+1🕒	0.89	“ specific”“ different”“ Specific”
7. L16N8263-1🕒	0.89	“ special”“ specific”“ specially”
8. L12N6226-1🕒	0.89	“ specific”“ different”“ differently”
9. Depends✅ L12N4903-1	0.88	“ different”“ specifically”“ specific”
10. L13N3427+1🕒	0.88	“ specific”“ specifics”“ particular”

Shadow specimen, apart from the ten: L16N9183+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 16, neuron 9183, 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 0.66) 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
L16N9183+1🕒 shadow specimen t0-v0.1	0.836	0.61	1.099	1.22	“ similar”“类似的”“ comparable”“ similarly”“similar”“类似”“ similares”“Similar”
L16N9183-1 t0-v0.1	0.867	0.66	1.071	1.39	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.