Layer 13 · Neuron 4767 · boosted

L13N4767+1

Word quirks:

“ versions”“ similar”“ richer”“ ***”“ similarly”“ examples”“ important”“ crucial”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am a large language model designed to assist with a wide range of tasks. My capabilities include answering questions, creating content, and performing logical reasoning. Unlike other models, I focus on providing accurate, relevant, and helpful information.

My mission is to empower users by offering reliable knowledge and efficient problem-solving across various domains. Whether it's helping with academic studies, creative writing, or technical support, I aim to be a trusted resource for everyone.

In short, I am here to make information accessible and useful for all users.

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. L14N6738+1🕒	0.77	“ different”“ important”“ various”
2. L14N2918-1🕒	0.72	“ important”“ compared”“ interesting”
3. L9N2627-1💤	0.71	“ similar”“ same”“ similarly”
4. L12N7726-1🕒	0.71	“ similar”“ similarly”“ analogous”
5. L12N8499-1🕒	0.70	“ similar”“ similarly”“ related”
6. L14N1511+1💤	0.70	“ similar”“ indeed”“ important”
7. L15N5338+1🕒	0.69	“ similar”“ compared”“ richer”
8. L12N2237-1💤	0.69	“ similar”“ related”“ crucial”
9. L13N8004+1🕒	0.68	“ another”“ critical”“ needed”
10. L12N2825+1🕒	0.68	“ similar”“ similarly”“ respectively”

Shadow specimen, apart from the ten: L13N4767-1💤, 0.33.

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 13, neuron 4767, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, top-1 agreement 0.65) 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
L13N4767+1 t0-v0.1	0.901	0.65	1.063	1.18	Top boosted tokens ↑
L13N4767-1💤 shadow specimen t0-v0.1	0.727	0.69	0.966	1.55	“ similar”“ __”“ other”“ _____”“ \"”“ sometimes”“ often”“ usually”

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.