Layer 19 · Neuron 8204 · boosted

L19N8204+1

Word quirks:

“对方”“对方的”“對方”“给对方”“让对方”“對方的”“ opposite”“ opposing”

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

Identity (according to the specimen itself)

I am Qwen3, a large-scale artificial intelligence developed by the other party. The other party's mission is to provide high-quality services for the other party. However, I must clarify that the other party is not the other party's product of the other party.

The other party's mission is to provide high-quality services for the other party. However, I must clarify that the other party is not the other party's product of the other party.

Let's rephrase:

Identity: I am an AI model developed by the other party.
Mission: To assist the other party in solving problems

The AI judge wasn't sure this holds together.

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. L30N4963+1💥	0.98	“对方”“对方的”“對方”
2. L22N7554+1🕒	0.88	“对方”“ Opp”“Opp”
3. L20N6809-1🕒	0.87	“另一方”“对方”“对面”
4. L17N7309-1🕒	0.82	“另一方”“ partner”“ counterpart”
5. L21N7699+1💤	0.78	“对面的”“对面”“ opposite”
6. L19N3317+1💤	0.69	“反面”“ opposite”“相反”
7. L21N2120-1🕒	0.67	“彼此”“彼此的”“各自”
8. L31N2853+1💥	0.67	“对”“對”“对用户”
9. L21N3976+1💤	0.67	“对他们”“對”“地对”
10. L25N3378-1🕒	0.67	“对”“對”“ terhadap”

Shadow specimen, apart from the ten: L19N8204-1💤, 0.26.

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 19, neuron 8204, 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
L19N8204+1 t0-v0.1	1.018	0.61	1.189	1.28	Top boosted tokens ↑
L19N8204-1💤 shadow specimen t0-v0.1	0.463	0.72	0.567	1.19	“ my”“aticon”“分类”“ distinctions”“的空间”“ My”“ distinction”“我曾”

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.