Layer 17 · Neuron 8834 · boosted

L17N8834+1

Word quirks:

“ (**”“ **'”“ **”“ answer”“;**”“ **«”“ **“”“ **#”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

My name is Qwen3.6. I am an AI assistant, my primary mission is to provide comprehensive, high-quality services to users. My core capabilities include:

My Mission Statement:

> "To be your AI assistant (Qwen3.6)"

Core Capabilities:

Answering Questions: Provide clear, accurate answers to your questions (Qwen3.6).
Document Summarization: Summarize documents for you (Qwen3.6).
Task Execution: Perform tasks such as **Qwen3.6

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. L16N7611+1🕒	0.96	“ ”“ "”“ **“”
2. L7N5407-1💤	0.95	“ "”“ ”“ "**”
3. L16N5304+1🕒	0.95	“ ”“ (”“ (**”
4. L13N508+1💤	0.94	“ ”“ #”“!**”
5. L16N553-1🕒	0.94	“ ”“ "”“ **“”
6. L15N4479+1💤	0.94	“ "”“ ”“ **“”
7. L17N7114-1🕒	0.94	“ ”“ .”“ (”
8. L17N1774-1💤	0.94	“ “”“;”“ **‘”
9. L7N7301+1💤	0.94	“ ”“ "”“ **“”
10. L9N559+1💤	0.94	“ ”“ "”“ **'”

Shadow specimen, apart from the ten: L17N8834-1🕒, -0.08.

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 8834, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L17N8834+1 t0-v0.1	1.583	0.54	1.800	1.10	Top boosted tokens ↑
L17N8834-1🕒 shadow specimen t0-v0.1	0.790	0.66	0.830	1.19	“ different”“ needed”“ differently”“就需要”“ __”“ needs”“ interesting”“需要了解”

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.