Layer 20 · Neuron 4312 · suppressed

L20N4312-1

Word quirks:

“agos”“肥”“肥料”“荷”“肥肉”“每期”“rega”“ priv”

🕒 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 Q (Qwen), a large language model developed by Alibaba Group. My core mission is to provide high-quality, efficient, and practical solutions for diverse scenarios by leveraging my training data and algorithmic updates. I aim to help users in areas such as:

Problem-solving: Breaking down complex issues into manageable steps with actionable strategies.
Content creation: Generating creative writing, technical documentation, or marketing materials tailored to specific needs.
Decision-making: Offering data-driven insights and logical reasoning to support informed choices.
Education: Explaining concepts from basic principles to advanced topics, including coding

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. L22N1286+1💤	0.78	“繁”“肥”“复制”
2. L28N3014-1🕒	0.78	“丰”“豐”“饱”
3. L25N9188+1🕒	0.78	“肥”“媒”“ drone”
4. L26N426-1💤	0.77	“肉”“汀”“泡脚”
5. L28N8655-1💤	0.77	“rag”“肥胖”“肥”
6. L26N1690+1🕒	0.76	“肥”“计量”“净”
7. L31N3360+1💥	0.76	“蒙”“肝”“脂”
8. L27N5790-1🕒	0.76	“金”“ гип”“石”
9. L28N2897+1💤	0.76	“肥”“čaj”“肤”
10. L20N5950+1🕒	0.76	“ fertil”“脂”“忘了”

Shadow specimen, apart from the ten: L20N4312+1🕒, -0.04.

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 4312, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.73) 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
L20N4312+1🕒 shadow specimen t0-v0.1	0.634	0.69	0.764	1.20	“ performed”“ Perform”“ performing”“perform”“ performs”“ perform”“Perform”“ Performing”
L20N4312-1 t0-v0.1	0.732	0.73	0.840	1.28	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.