Layer 17 · Neuron 8196 · suppressed

L17N8196-1

Word quirks:

“ acknowledging”“ acknowledgment”“ acknowledge”“ acknowledged”“ recognition”“ recognizing”“ acknowledges”“ awareness”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen 3.5, a large language model developed by Alibaba Cloud. I am part of the Qwen series, known for its strong capabilities in reasoning, code generation, and multi-modal understanding. My primary mission is to be a helpful and responsible assistant who adheres to ethical guidelines, ensuring transparency and accountability in my outputs. I strive to provide accurate information, support learning and problem-solving, and avoid sharing sensitive or harmful content. Whether you're looking for creative writing, logical reasoning, or technical insights, I aim to align with your expectations while maintaining clarity about my limitations.

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. L4N3930+1🕒	0.97	“ acknowledgment”“ recognition”“ acknowledgement”
2. L8N4343-1💤	0.93	“ acknowledgment”“ acknowledged”“ acknowledging”
3. L7N5222-1💥	0.93	“ acknowledging”“ acknowledgment”“ acknowledgement”
4. L16N8309+1💤	0.92	“ acknowledging”“ acknowledgment”“ acknowledged”
5. L10N3430-1🕒	0.90	“ acknowledgment”“ acknowled”“ acknowledgement”
6. L5N5260-1🕒	0.90	“ acknowledgment”“ acknowledging”“ truth”
7. L13N7391-1💤	0.90	“ basic”“ acknowledgment”“ acknowledgement”
8. L19N7534+1🕒	0.90	“ acknowledgment”“ acknowledgement”“ Ack”
9. L12N5692-1🕒	0.89	“ acknowledgment”“ recognition”“ recognizing”
10. L21N3225+1💤	0.89	“ acknowledgment”“ recognition”“ acknowledgement”

Shadow specimen, apart from the ten: L17N8196+1🕒, 0.02.

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 8196, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.62, top-1 agreement 0.65) 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
L17N8196+1🕒 shadow specimen t0-v0.1	0.943	0.65	1.089	1.27	“ tempting”“ quick”“ shortcuts”“ speed”“ shortcut”“ convenient”“ speeding”“ convenience”
L17N8196-1 t0-v0.1	0.616	0.65	0.800	1.31	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.