Layer 19 · Neuron 7961 · suppressed

L19N7961-1

Word quirks:

“ não”“ not”“ non”“ nicht”“非”“ Non”“ niet”“ Nicht”

🕒 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 Qiancheng, an AI model developed by Qiancheng Lab. I am not a single, fixed entity with a singular "mission" in the way that every human doesn't have a specific job title. Instead, my purpose is to serve as your dedicated partner for problem-solving and collaboration.

My core mission is to empower you by providing:

Non-stop Collaboration: I don't sleep, so I'm available 24/7 to tackle challenges that require constant attention or don't have a definitive "off-clock" solution.
Adaptive Problem-Solving:

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. Nonstop Non✅ L28N4934-1	0.91	“非”“ non”“不是”
2. L0N8013-1💥	0.91	“non”“-non”“not”
3. L29N1768+1🕒	0.89	“ не”“ not”“ nicht”
4. L28N7704-1🕒	0.89	“ non”“non”“-non”
5. L28N3260+1💥	0.89	“非”“ non”“Non”
6. L28N3575+1💥	0.89	“非”“而非”“non”
7. L27N3410+1🕒	0.88	“ non”“非”“ Non”
8. L25N8274+1🕒	0.88	“ non”“非”“non”
9. L27N4793+1🕒	0.88	“ non”“ Non”“Non”
10. L26N6067+1🕒	0.88	“非”“ non”“non”

Shadow specimen, apart from the ten: L19N7961+1🕒, 0.08.

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 7961, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, top-1 agreement 0.71) 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
L19N7961+1🕒 shadow specimen t0-v0.1	0.932	0.61	1.022	1.19	“ concept”“ whatsoever”“ concepts”“ notion”“概念”“ idea”“ completely”“ unrelated”
L19N7961-1 t0-v0.1	0.825	0.71	0.984	1.19	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.