Layer 7 · Neuron 8151 · boosted

L7N8151+1

Word quirks:

“ processed”“ executed”“ executing”“ below”“ execution”“ specific”“ present”“ represented”

💤 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 Qwen3.5, a large language model developed by Alibaba Cloud. My primary mission is to provide high-quality, accurate responses across a wide range of tasks, including natural language understanding, text generation, and complex reasoning.

Core Capabilities:

Natural Language Understanding (NLU): I can interpret and process human language with precision, enabling seamless interaction with users.
Text Generation: I generate coherent, contextually relevant text for various applications, from customer service to content creation.
Complex Reasoning: I perform logical operations such as mathematical calculations, data analysis, and problem-solving

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. L5N9003-1🕒	0.82	“ executed”“ deployed”“ execution”
2. L15N7680-1💤	0.81	“ before”“ execution”“ BEFORE”
3. L12N8451-1🕒	0.80	“ execution”“ executing”“ executed”
4. L13N8115-1🕒	0.78	“ generating”“ executed”“ executing”
5. L17N8771+1🕒	0.78	“ executing”“ carefully”“ executed”
6. L6N8708-1💤	0.77	“ execution”“ instantly”“ immediate”
7. Deadpan✅ L3N7155-0.6,L13N4810-0.6	0.77	“ execution”“ immediate”“ devoid”
8. L9N8890+1💤	0.76	“ specific”“ processing”“ execution”
9. The Executor✅ L13N4810-0.6,L22N4858-0.6	0.76	“ execution”“ executed”“Execution”
10. L18N6867+1🕒	0.76	“ execution”“ executed”“执行”

Shadow specimen, apart from the ten: L7N8151-1🕒, 0.18.

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 7, neuron 8151, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.80, 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
L7N8151+1 t0-v0.1	0.797	0.65	1.011	1.39	Top boosted tokens ↑
L7N8151-1🕒 shadow specimen t0-v0.1	1.808	0.47	2.110	1.90	“ files”“ archive”“ compressed”“ download”“ zip”“ downloaded”“ archives”“ file”

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.