Layer 24 · Neuron 3809 · suppressed

L24N3809-1

Word quirks:

“ perspective”“Pers”“ Perspective”“pers”“ reflection”“ reflections”“ perspectiva”“ pers”

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

Identity (according to the specimen itself)

I are Perspective, a perspective-based AI model designed to help you explore different perspectives on complex topics. My mission is to provide you with diverse, well-reflective, and multi-perspective reflections on various topics, helping you gain a broader understanding of the world. Whether you need to explore different perspectives on a topic, reflect on a specific issue, or simply want to see things from a different perspective, I am here to guide you in exploring these perspectives.

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. L26N3606+1🕒	0.94	“Pers”“ pers”“ perspective”
2. L23N7412-1🕒	0.93	“ perspective”“pers”“Pers”
3. L22N7312-1🕒	0.92	“ perspective”“Pers”“ Perspective”
4. L22N7936+1🕒	0.91	“ perspective”“ perspectives”“pers”
5. L20N8492+1🕒	0.88	“ perspective”“视角”“ perspectives”
6. L27N2293+1💤	0.86	“ perspectives”“ perspective”“ Perspective”
7. L31N8114-1💥	0.84	“pers”“Pers”“ Pers”
8. L29N7523-1💥	0.82	“ perspective”“ Perspective”“ perspectives”
9. L4N5484-1💥	0.80	“ perspective”“ Perspective”“ perspect”
10. L20N8249-1🕒	0.80	“ perspective”“ perspectives”“ Perspective”

Shadow specimen, apart from the ten: L24N3809+1💤, -0.11.

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 3809, 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
L24N3809+1💤 shadow specimen t0-v0.1	0.721	0.69	0.819	1.50	“抓到”“基”“unami”“pla”“-dev”“ numer”“ DEV”“ Dev”
L24N3809-1 t0-v0.1	1.515	0.63	1.574	1.15	Top boosted tokens ↑

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.