Layer 11 · Neuron 2578 · boosted

L11N2578+1

Word quirks:

“ you”“ your”“ here”“ usually”“ yours”“ now”“ implied”“ yourself”

🕒 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 Qwen3.5, the large language model developed by Alibaba Cloud. My mission is to provide you with a seamless experience in interacting with complex problems, such as analyzing mathematical expressions or reasoning about code. For example, if you want to find the solution to a specific problem, I can do so without needing to explain the process.

To further clarify my capabilities, here are some key aspects of my design:

Multilingual Support: I support over 100 languages, including Chinese, which means I can communicate with users from various regions and cultures.
Code Generation: I can generate code

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. L11N8729-1💤	0.93	“ here”“ your”“ often”
2. L14N1954-1🕒	0.93	“ now”“ here”“ your”
3. L9N7598-1💤	0.92	“ now”“ your”“ here”
4. L9N4862-1💤	0.92	“ your”“ yours”“ you”
5. L7N8061-1🕒	0.92	“ you”“ your”“ here”
6. L8N5041-1💤	0.92	“ your”“ now”“ you”
7. L13N8561-1🕒	0.92	“ your”“ yourself”“ you”
8. L8N6188-1💤	0.91	“ your”“ now”“ my”
9. L15N1448+1💤	0.91	“ your”“ here”“ you”
10. L10N7135+1🕒	0.91	“ your”“ now”“ here”

Shadow specimen, apart from the ten: L11N2578-1🕒, -0.17.

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 11, neuron 2578, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, top-1 agreement 0.65) 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
L11N2578+1 t0-v0.1	0.967	0.65	1.330	1.41	Top boosted tokens ↑
L11N2578-1🕒 shadow specimen t0-v0.1	1.047	0.61	1.276	1.39	“ "”“ "”“ "”“ “”“ "*”“ “”“ ("”“—"”

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.