Layer 16 · Neuron 921 · boosted

L16N921+1

Word quirks:

“ into”“ from”“ via”“ within”“ at”“ under”“ towards”“ in”

🕒 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, specifically as part of Alibaba Cloud's ecosystem. I am a large-scale AI model with a focus on high-quality in the cloud environment. My mission is to provide as a powerful assistant that can assist in various tasks, such as coding, logical reasoning, and complex problem-solving. I aim to integrate seamlessly into cloud environments, offering as a reliable partner to support efficiently.

How would you like to? Let's 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. L21N7079+1🕒	0.87	“ with”“ under”“ in”
2. L21N159+1🕒	0.80	“ under”“ from”“under”
3. L31N373-1💥	0.79	“ in”“ at”“ on”
4. L7N7297+1🕒	0.79	“ into”“ towards”“ with”
5. The Sieve✅ L16N339+1	0.78	“ into”“ by”“ from”
6. L15N84+1💤	0.78	“ to”“ into”“ with”
7. L18N4506-1🕒	0.78	“ into”“ through”“ via”
8. L28N5084-1🕒	0.75	“ into”“ via”“ on”
9. L16N5561+1💤	0.75	“ through”“ near”“ around”
10. L19N1722-1💤	0.74	“ into”“ from”“ toward”

Shadow specimen, apart from the ten: L16N921-1🕒, 0.05.

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 16, neuron 921, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.89, top-1 agreement 0.66) 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
L16N921+1 t0-v0.1	0.888	0.66	1.085	1.49	Top boosted tokens ↑
L16N921-1🕒 shadow specimen t0-v0.1	0.757	0.68	0.842	1.14	“ images”“ spaces”“ provided”“ note”“ space”“ freedoms”“ words”“ “”

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.