Layer 25 · Neuron 3511 · boosted

L25N3511+1

Word quirks:

“onton”“砂糖”“米”“靶”“三无”“ród”“rees”“原则”

🕒 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 N**eoadvisor, a high-performance AI assistant designed to help you make better decisions with confidence. My mission is to serve as your personal strategic advisor, providing clear, actionable, and context-aware guidance across a wide range of domains.

Here’s how I can help:

Make Better Decisions: Analyze complex situations, weigh pros and cons, and offer balanced recommendations tailored to your goals.
Solve Problems: Tackle challenges in business, personal life, or technical tasks with structured thinking and creative solutions.
Learn and Grow: Explain concepts in simple terms, guide you

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. L28N5348-1💤	0.80	“кол”“لاف”“默”
2. L28N3156+1💤	0.79	“谷”“素”“素的”
3. L26N5941+1🕒	0.79	“宗主”“红”“米”
4. L27N6500+1💤	0.79	“番”“米的”“荧”
5. L28N7059-1💤	0.78	“新房”“汤”“镜”
6. L28N2686+1💤	0.78	“回”“首”“onz”
7. L27N6384-1💤	0.78	“角”“纯”“ Kont”
8. L26N3205+1💤	0.77	“吃”“ogat”“靶”
9. L30N6718+1💤	0.77	“货”“府”“墨”
10. L26N8298-1💤	0.77	“仕”“商”“坚毅”

Shadow specimen, apart from the ten: L25N3511-1🕒, 0.57.

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 25, neuron 3511, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.66, top-1 agreement 0.72) 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
L25N3511+1 t0-v0.1	0.664	0.72	0.713	1.30	Top boosted tokens ↑
L25N3511-1🕒 shadow specimen t0-v0.1	1.512	0.56	1.532	1.30	“絲”“ Silk”“uin”“ymp”“シル”“ placements”“够了”“丝”

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.