Layer 10 · Neuron 6203 · suppressed

L10N6203-1

Word quirks:

“ carefully”“ special”“ specially”“ your”“ careful”“ typically”“ reports”“ concern”

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

Identity (according to the specimen itself)

I am a large language model designed to assist with complex tasks. My training data is vast and diverse, containing information from many sources. I have specific instructions that help me process this information accurately and efficiently. My goal is to provide high-quality responses while maintaining clarity and precision in the output.

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. L10N866-1🕒	0.81	“ carefully”“ usually”“ specially”
2. L10N7965+1🕒	0.79	“ carefully”“ usually”“ careful”
3. L12N6774-1💤	0.79	“ specially”“ carefully”“ deliberately”
4. L11N4393-1🕒	0.78	“ carefully”“ careful”“ focused”
5. L11N8475-1🕒	0.76	“ specially”“ specifically”“ ONLY”
6. L16N607+1🕒	0.76	“ general”“ special”“ carefully”
7. L13N8714+1🕒	0.75	“ carefully”“ constantly”“ careful”
8. L11N6963-1🕒	0.74	“ carefully”“ properly”“ balance”
9. L12N1479+1💤	0.74	“ thinking”“ always”“ your”
10. L9N8994+1🕒	0.74	“ carefully”“ careful”“ now”

Shadow specimen, apart from the ten: L10N6203+1🕒, -0.10.

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 6203, 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
L10N6203+1🕒 shadow specimen t0-v0.1	1.490	0.54	1.750	1.42	“<\|im_end\|>”“⏎”“ ____”“ _____”“ ______”“ -->”“ login”“ users”
L10N6203-1 t0-v0.1	1.127	0.57	1.360	1.42	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.