Layer 20 · Neuron 3336 · boosted

L20N3336+1

Word quirks:

“实施”“ implementation”“ employing”“ implement”“ implementing”“实施了”“ utilizing”“ enable”

🕒 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 Qwen, a large-scale model developed by Tencent's Tencent Enabling AI (Tencent AI for Health). My mission is to leverage advanced artificial intelligence to facilitate intelligent decision-making, enhance user experience, and promote efficient problem-solving across various scenarios. I aim to provide accurate, context-aware, and ethically grounded responses while supporting tasks like multi-step reasoning, content creation, and adaptive planning. Whether through precise question answering, logical analysis, or creative collaboration, I strive to deliver robust solutions that align with the goals of users and the broader ecosystem of AI-driven innovation.

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. L26N1644-1🕒	0.91	“实施”“ implement”“ implements”
2. L25N3352-1💤	0.90	“实施”“实施了”“implementation”
3. L30N1992+1💥	0.90	“实施”“ implementation”“ implement”
4. L24N1449-1💥	0.89	“plement”“实施”“ Implement”
5. L27N8867-1💤	0.89	“ implementing”“ implemented”“ execution”
6. L22N7149+1🕒	0.89	“ implementation”“ implementations”“ implementing”
7. L30N4762+1💥	0.88	“实施”“施行”“实施的”
8. L25N4193+1🕒	0.88	“ Implement”“实施”“ implement”
9. L24N4382+1🕒	0.88	“ implement”“ implementing”“ Implement”
10. L24N8417-1💤	0.87	“实施”“implementation”“Implement”

Shadow specimen, apart from the ten: L20N3336-1🕒, 0.03.

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 20, neuron 3336, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.93, top-1 agreement 0.61) 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
L20N3336+1 t0-v0.1	0.929	0.61	0.995	1.20	Top boosted tokens ↑
L20N3336-1🕒 shadow specimen t0-v0.1	0.702	0.71	0.872	1.22	“残余”“ “”“ reminds”“ defined”“ leftover”“ Rem”“ remaining”“定義”

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.