Layer 15 · Neuron 9054 · boosted

L15N9054+1

Word quirks:

“ dedicated”“ special”“ rapid”“ specially”“ rapidly”“ fast”“ _____”“ required”

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

Identity (according to the specimen itself)

Hello! I am Qwen3.5, a large-scale language model developed by Alibaba Cloud. My mission is to provide high-quality, efficient, and intelligent AI solutions that can serve diverse user needs across various domains.

Key Features & Capabilities:

Advanced Language Understanding: I excel at natural language processing, capable of handling complex tasks with precision.
Full-Stack Development Support: My development capabilities cover the entire spectrum, from conceptualization to deployment.
Visual Analysis: I can interpret charts, graphs, and diagrams effectively.
Agentic Workflow: I support autonomous task management,

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. L9N4776-1🕒	0.77	“ specially”“ --”“ specialised”
2. L9N4951-1🕒	0.76	“ special”“ faster”“ specially”
3. L12N6725+1🕒	0.73	“ slower”“ speeds”“ specific”
4. L11N5640-1🕒	0.72	“ speed”“ faster”“ optimized”
5. L16N7095+1💤	0.71	“ rapid”“ rapidly”“ fast”
6. L16N7430+1🕒	0.71	“ quick”“ simplified”“ forgotten”
7. L11N7695-1🕒	0.70	“ quickly”“ rapidly”“ quick”
8. L13N3930-1🕒	0.70	“ instantly”“ quicker”“ faster”
9. L4N7867-1🕒	0.70	“ quick”“ quicker”“ swift”
10. L19N7630-1🕒	0.70	“ rapid”“ rapidly”“ pace”

Shadow specimen, apart from the ten: L15N9054-1🕒, 0.01.

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 9054, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.63, top-1 agreement 0.70) 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
L15N9054+1 t0-v0.1	0.628	0.70	0.764	1.25	Top boosted tokens ↑
L15N9054-1🕒 shadow specimen t0-v0.1	0.769	0.69	1.029	1.23	“ --”“ code”“ info”“ ---”“ ——”“ ____”“ —”“ __”

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.