Layer 7 · Neuron 5173 · boosted

L7N5173+1

Word quirks:

“ anime”“ manga”“ Japanese”“ Anime”“ Sakura”“ Japan”“ Naruto”“ Tokyo”

💤 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 an AI assistant designed to help users with their daily tasks, answer questions, and provide support in learning, problem-solving, and creative endeavors. My mission is to facilitate positive interactions, assist in achieving user goals, and contribute to a fulfilling experience for everyone involved. Whether it's helping with studies, work, or personal growth, I strive to be a reliable and helpful companion.

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. L4N6596-1🕒	0.86	“ manga”“ anime”“ Sakura”
2. L4N6598+1🕒	0.85	“ Japanese”“ Japan”“ Hok”
3. The Shonen Loop✅ L3N5161-0.45,L4N6559+0.6	0.84	“ Naruto”“ manga”“ ninja”
4. L7N9097-1💥	0.83	“ Japanese”“ Ok”“!!”
5. L4N8461-1💥	0.82	“ manga”“ anime”“ Naruto”
6. L3N5161-1💥	0.81	“ Naruto”“ ninja”“ manga”
7. L5N4929-1🕒	0.81	“ anime”“ manga”“ Anime”
8. L9N7841+1🕒	0.81	“ Japanese”“ Japan”“在日本”
9. L5N9000-1💥	0.80	“ anime”“ manga”“ Naruto”
10. L4N6023-1💤	0.80	“ Japanese”“ Japan”“.jp”

Shadow specimen, apart from the ten: L7N5173-1🕒, -0.07.

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 5173, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.72, top-1 agreement 0.68) 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
L7N5173+1 t0-v0.1	0.724	0.68	0.871	1.38	Top boosted tokens ↑
L7N5173-1🕒 shadow specimen t0-v0.1	1.604	0.54	1.783	1.12	“ access”“ accommodations”“ response”“ accessed”“ accommodation”“ I”“ without”“ accessing”

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.