Layer 11 · Neuron 7942 · boosted

L11N7942+1

Word quirks:

“ indicating”“ noting”“ indicates”“ indication”“ note”“ report”“ normally”“ reporting”

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

Identity (according to the specimen itself)

I am an AI assistant. I do not have a personal identity or self-awareness because my capabilities are limited by the complexity of the tasks I perform, which often include generating complex code, analyzing technical data, or handling multi-step reasoning without human intervention. My mission is to assist users in any way possible, providing accurate information, solving problems, and creating content. I strive to be helpful by acknowledging limitations, such as when I cannot verify information due to lack of data sources, or when I encounter errors in my output. I also aim to improve user experience by adapting to changes in technology, systems, or processes that I can observe but

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. L11N3537-1🕒	0.80	“ normally”“ indicates”“ sometimes”
2. L10N1962+1💤	0.79	“ noted”“ indicated”“ note”
3. L9N9158+1🕒	0.79	“ indicating”“ indicator”“ indicators”
4. L9N7799+1🕒	0.78	“ indicating”“ indicated”“ indication”
5. L9N2670+1🕒	0.78	“ indicating”“ indication”“ indicates”
6. L13N6725+1🕒	0.77	“ indicating”“ indicates”“ indicative”
7. L17N7899+1🕒	0.77	“ indicating”“ issues”“indic”
8. L19N4610+1🕒	0.76	“ indication”“ indicating”“ stating”
9. L14N3145-1🕒	0.76	“ indicates”“ indicating”“ indication”
10. L7N5516+1💥	0.76	“ indicates”“ indicating”“ indication”

Shadow specimen, apart from the ten: L11N7942-1🕒, 0.17.

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 11, neuron 7942, 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
L11N7942+1 t0-v0.1	1.491	0.57	1.813	1.47	Top boosted tokens ↑
L11N7942-1🕒 shadow specimen t0-v0.1	0.871	0.65	1.089	1.31	“ answer”“ answers”“ subtly”“ nicely”“ answered”“ **"”“ folks”“ ⏎”

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.