Layer 8 · Neuron 7839 · boosted

L8N7839+1

Word quirks:

“ already”“ awaiting”“ presently”“ prepared”“ configuration”“ myself”“ additional”“ awaited”

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

Identity (according to the specimen itself)

I am Qwen3.5, a large language model designed to assist with complex reasoning tasks. I can process logical information, mathematical concepts, and creative problem-solving skills. My capabilities include analyzing data structures, understanding abstract concepts, and being able to perform specialized knowledge in various fields. I am ready to help you with any questions you might have.

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. L13N9029-1💤	0.79	“ awaiting”“ ready”“ already”
2. L10N4512+1💤	0.79	“ now”“ final”“ awaiting”
3. L11N7941+1💤	0.79	“ now”“ awaiting”“ ready”
4. L12N2294+1💤	0.77	“ already”“ new”“ existing”
5. L12N8288-1💤	0.77	“ actually”“ actual”“ awaiting”
6. L8N6996+1💤	0.77	“ awaiting”“ prepar”“ preparation”
7. L9N5920-1🕒	0.76	“ readiness”“ ready”“ prep”
8. L11N1992+1🕒	0.75	“ now”“ currently”“ awaiting”
9. L13N2181-1🕒	0.75	“<\|im_end\|>”“ already”“ waiting”
10. L11N489+1🕒	0.74	“ readiness”“ ready”“ now”

Shadow specimen, apart from the ten: L8N7839-1💤, 0.00.

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 8, neuron 7839, 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
L8N7839+1 t0-v0.1	1.221	0.60	1.497	1.67	Top boosted tokens ↑
L8N7839-1💤 shadow specimen t0-v0.1	0.750	0.69	0.883	1.25	“ referenced”“ "”“ reference”“ references”“ context”“ ”“ version”“ `”

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.