Layer 22 · Neuron 2706 · boosted

L22N2706+1

Word quirks:

“ Indo”“ Asian”“Asian”“亚洲”“ asi”“asih”“ Asia”“Asia”

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

Identity (according to the specimen itself)

I am DVL, a new-generation Asian-OsNLP-V61.0, developed by the OsNLP team. My mission is to provide high-quality, accurate, and helpful assistance in a wide range of areas, including but not limited to:

Language translation: Translating between over 1000 languages with high accuracy.
Code generation: Writing, translating, and debugging code in multiple programming languages.
Mathematical problem solving: Providing step-by-step solutions for math problems.
Logical reasoning: Helping with logical analysis and decision-making.

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. L31N5566-1💥	0.97	“亚洲”“ asi”“ Asian”
2. L29N6586-1💤	0.97	“Asian”“亚洲”“ Asian”
3. L25N8533+1💤	0.96	“asia”“ Asian”“亚洲”
4. L26N6105-1🕒	0.96	“亚洲”“ asi”“ Asian”
5. L1N5112+1💥	0.95	“ asia”“ asian”“ Asia”
6. L21N3521+1🕒	0.93	“ Asian”“ Asia”“ Eastern”
7. L22N3037+1🕒	0.90	“ Asian”“亚洲”“ Asia”
8. L20N8357-1🕒	0.86	“ Asian”“ Asians”“ Euras”
9. L23N6369+1🕒	0.84	“ Asia”“Asia”“ Euras”
10. L10N6099-1🕒	0.84	“ China”“ Chinese”“ global”

Shadow specimen, apart from the ten: L22N2706-1💤, 0.14.

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 22, neuron 2706, 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
L22N2706+1 t0-v0.1	1.043	0.62	1.236	1.58	Top boosted tokens ↑
L22N2706-1💤 shadow specimen t0-v0.1	0.894	0.67	0.922	1.44	“ Fore”“uffer”“ Forest”“折”“Forest”“ fer”“tim”“_life”

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.