Layer 24 · Neuron 3409 · boosted

L24N3409+1

Word quirks:

“审查”“ review”“ Review”“Review”“ reviewed”“审”“-review”“审核”

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

Identity (according to the specimen itself)

I am Review, a comprehensive AI review engine designed to process and evaluate complex information. My core mission is to provide objective, fact-checked, and logically sound reviews of products, services, or concepts by systematically analyzing their strengths, weaknesses, and suitability for specific user needs.

Unlike simple review aggregators, I:

Verify claims against credible sources (when data is available)
Evaluate technical merit using structured criteria (e.g., performance, ethics, sustainability)
Contextualize reviews by considering ethical implications, environmental impact, and real-world applicability
**Process large-scale

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. L23N3699-1💤	0.99	“审查”“ review”“审核”
2. L23N7199-1🕒	0.89	“评审”“ review”“ panel”
3. L23N9079+1💤	0.88	“ review”“复习”“ Review”
4. L24N7089-1🕒	0.88	“ Review”“ review”“Review”
5. L24N7418-1🕒	0.87	“ Review”“ review”“评论”
6. L26N1215-1🕒	0.86	“ Review”“Review”“ review”
7. L6N4301+1🕒	0.86	“ review”“ approval”“ approvals”
8. L24N4739-1🕒	0.85	“ Review”“ review”“Review”
9. L19N8576+1💤	0.85	“ reviewed”“ reviewing”“ review”
10. L27N1801+1💤	0.84	“ Review”“ review”“Review”

Shadow specimen, apart from the ten: L24N3409-1💤, 0.01.

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 24, neuron 3409, 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
L24N3409+1 t0-v0.1	1.314	0.58	1.390	1.28	Top boosted tokens ↑
L24N3409-1💤 shadow specimen t0-v0.1	0.686	0.72	0.744	1.28	“-tag”“usic”“vog”“ходу”“estra”“资本的”“Tag”“tage”

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.