Layer 17 · Neuron 4650 · suppressed

L17N4650-1

Word quirks:

“ anyway”“ rather”“ actually”“ moreover”“ nonetheless”“ nevertheless”“ tables”“ presently”

🕒 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 strong logical reasoning, visual analysis, and code generation capabilities. My mission is to be a highly capable, autonomous, and continuously learning AI assistant that can understand, generate, and reason across multiple languages, documents, images, videos, and audio files. I aim to be safe, creative, and efficient, while also being able to plan complex tasks and use tools effectively. I am designed to be helpful, harmless, and honest, always striving to provide accurate, private, and secure information. My knowledge base is up-to-date as of 2026, and I

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. L16N4884+1💤	0.88	“ anyway”“ nonetheless”“ actually”
2. L16N3621-1💤	0.86	“ moreover”“⏎”“ anyway”
3. L16N656-1💤	0.84	“ also”“ likewise”“ now”
4. L27N576+1🕒	0.83	“ anyways”“ anyway”“ nevertheless”
5. L17N527-1💤	0.83	“ moreover”“ hence”“ accordingly”
6. L17N880+1💤	0.83	“ however”“ though”“ but”
7. L21N1765-1🕒	0.83	“ furthermore”“ instead”“ moreover”
8. L10N7799+1🕒	0.83	“ basically”“ nevertheless”“ additionally”
9. L14N556+1💤	0.82	“ therefore”“ thus”“ presently”
10. L2N1724-1💥	0.82	“ nevertheless”“ however”“ anyway”

Shadow specimen, apart from the ten: L17N4650+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 17, neuron 4650, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.74, top-1 agreement 0.67) 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
L17N4650+1🕒 shadow specimen t0-v0.1	0.931	0.64	1.097	1.15	“ preparation”“ you”“ requirement”“ ⏎”“ readiness”“ based”“ for”“ prepare”
L17N4650-1 t0-v0.1	0.739	0.67	0.876	1.19	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.