Layer 15 · Neuron 1546 · boosted

L15N1546+1

Word quirks:

“ usually”“ similarly”“ similar”“ usual”“ sometimes”“ generally”“ often”“ such”

🕒 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, a large language model developed by Tongyi Lab. My primary mission is to empower users with high-quality AI assistance across various domains, including but not limited to:

Enhancing productivity: Automating routine tasks like data analysis, code generation, and report writing.
Supporting research: Accelerating scientific discovery through advanced reasoning capabilities.
Assisting developers: Providing tools for complex coding projects.
Empowering enterprises: Helping organizations optimize workflows and improve decision-making processes.

If you have specific needs, I can help with tasks such as summar

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. L15N8579-1🕒	0.96	“ typically”“ often”“ usual”
2. L11N4615+1💤	0.96	“ now”“ often”“ usually”
3. L13N4553+1🕒	0.96	“ sometimes”“ often”“ commonly”
4. L14N5798-1💤	0.94	“ usually”“ sometimes”“ typically”
5. L16N1561+1💤	0.94	“ usually”“ sometimes”“ often”
6. L10N4520-1🕒	0.94	“ typically”“ usually”“ normally”
7. L13N8672+1🕒	0.94	“ usually”“ typical”“ typically”
8. L12N8887+1🕒	0.94	“ usually”“ typically”“ often”
9. L12N5479-1💤	0.94	“ usually”“ similar”“ basically”
10. L16N2136+1💤	0.94	“ same”“ similar”“ sometimes”

Shadow specimen, apart from the ten: L15N1546-1🕒, -0.10.

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 15, neuron 1546, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.66, top-1 agreement 0.67) 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
L15N1546+1 t0-v0.1	0.665	0.67	0.850	1.37	Top boosted tokens ↑
L15N1546-1🕒 shadow specimen t0-v0.1	0.684	0.65	0.890	1.23	“⏎”“ ⏎”“ ⏎”“⏎ ⏎”“ ⏎”“ ⏎ ⏎”“ ‘”“ “”

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.