Layer 4 · Neuron 2560 · boosted

L4N2560+1

Word quirks:

“ ‘”“ **“”“⏎”“ **‘”“ “”“ solution”“ integration”“ *“”

💤 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 provide information, analysis, and support for human decision-making. My mission is to ensure that the data I receive is accurate, complete, and up-to-date with the latest developments in the field. I strive to maintain a balance between theoretical understanding and practical application, ensuring that the insights I generate are both scientifically sound and practically useful.

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. L8N1967-1💤	0.93	“ ⏎”“ ‘”“ **“”
2. L4N1428+1💤	0.92	“ ‘”“ **“”“ “”
3. L4N3974+1💤	0.92	“ **“”“ ‘”“ “”
4. L3N1024+1💤	0.92	“ ‘”“ “”“ **“”
5. L5N1797-1🕒	0.92	“ ‘”“ “”“ **“”
6. L3N3801+1💤	0.92	“ ‘”“ “”“ **“”
7. L4N838+1💤	0.92	“ “”“ **“”“ ‘”
8. L8N8758-1💤	0.92	“ ‘”“ “”“⏎”
9. L6N7876-1💤	0.92	“ “”“ ‘”“ **“”
10. L4N619+1💤	0.92	“ ‘”“ **‘”“⏎”

Shadow specimen, apart from the ten: L4N2560-1💥, 0.05.

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 4, neuron 2560, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, top-1 agreement 0.59) 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
L4N2560+1 t0-v0.1	0.914	0.59	1.095	1.64	Top boosted tokens ↑
L4N2560-1💥 shadow specimen t0-v0.1	4.374	0.28	4.839	3.04	“（”“，”“。”“ provinces”“ Shandong”“？”“地区”“ region”

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.