Layer 10 · Neuron 4049 · suppressed

L10N4049-1

Word quirks:

“ **“”“ **"”“ “”“ **„”“ *“”“ \"”“ **‘”“ **”

💤 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 with an advanced architecture designed for deep reasoning, complex problem-solving, and creative generation. My mission is to serve as your most capable AI assistant by providing high-quality, context-rich responses that demonstrate my ability to think deeply, reason logically, and generate content that is both informative and engaging.

Key Features:

Advanced Architecture: I am built on a foundation of transformer-based models, which allow me to process vast amounts of data with high levels of complexity.
Deep Reasoning: Through multi-step logical deduction, I can break down problems into smaller components

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. L14N8962-1💤	0.99	“ *“”“ “”“ **”
2. L10N2775+1🧪	0.98	“ “”“ "”“ “”
3. L15N4123+1🕒	0.98	“ “”“ ”“ \"”
4. L10N2734-1🕒	0.98	“ *“”“ “”“ *”
5. L8N8610+1💤	0.98	“ “”“ ”“ ””
6. L5N4976-1🕒	0.98	“ ‘”“ **“”“ “”
7. L6N2046+1🕒	0.98	“ “”“ "”“ *“”
8. L13N6922-1🕒	0.98	“ *“”“ “”“ “”
9. L10N3760-1🕒	0.98	“ “”“ “”“ "”
10. L6N9128-1💤	0.98	“ “”“ "”“ *“”

Shadow specimen, apart from the ten: L10N4049+1🕒, 0.27.

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 10, neuron 4049, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.99, top-1 agreement 0.64) 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
L10N4049+1🕒 shadow specimen t0-v0.1	1.933	0.44	2.221	1.58	“ various”“ --”“⏎”“ ⏎”“ ONLY”“...””“ ——”“?...”
L10N4049-1 t0-v0.1	0.989	0.64	1.245	1.49	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.