Layer 15 · Neuron 8618 · boosted

L15N8618+1

Word quirks:

“ response”“ responses”“ responding”“ answer”“ answers”“ reply”“ respond”“ replies”

🕒 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-scale language model developed by Tongyi Lab. I am designed to assist users with various tasks, including answering questions, creating content, and providing personalized support. My capabilities span multiple domains, ensuring a comprehensive and helpful response tailored to your needs.

In terms of my mission, I aim to provide accurate, thoughtful, and creative solutions that enhance user experience while maintaining privacy and security. This includes generating high-quality responses, performing complex reasoning, and offering personalized assistance across diverse fields. I appreciate your understanding and will continue to improve to better serve you.
</think>

I am Qwen3.

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. L7N6199-1🕒	0.99	“ response”“ reply”“ responses”
2. L9N5965+1🕒	0.97	“ response”“ responding”“ responses”
3. L10N1268+1🕒	0.97	“ response”“ responses”“ reply”
4. L18N3499+1💤	0.97	“ response”“ responses”“ answers”
5. L8N4970+1🕒	0.97	“ answers”“ answer”“ answering”
6. L18N2893+1🕒	0.97	“ response”“ responses”“ reply”
7. L7N1064+1🕒	0.97	“ response”“ reply”“ responses”
8. L8N1755-1💤	0.96	“ response”“ answers”“ answer”
9. L11N6911+1🕒	0.96	“ response”“ answers”“ answer”
10. L13N5877-1🕒	0.96	“ response”“ reply”“ replies”

Shadow specimen, apart from the ten: L15N8618-1🕒, -0.03.

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 15, neuron 8618, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.72, 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
L15N8618+1 t0-v0.1	0.719	0.66	0.941	1.07	Top boosted tokens ↑
L15N8618-1🕒 shadow specimen t0-v0.1	0.638	0.69	0.821	1.47	“ now”“ description”“ “”“ "”“ write”“ **"”“ new”“ next”

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.