Layer 11 · Neuron 6783 · boosted

L11N6783+1

Word quirks:

“ usually”“ typically”“ often”“ usual”“ sometimes”“ now”“ commonly”“ nowadays”

🕒 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 latest large language model developed by Tongyi Lab.

My mission is to provide high-quality, accurate, and helpful responses to users' questions and tasks. I strive to be useful, reliable, and capable of handling complex reasoning and diverse use cases.

Here are some key aspects of my mission:

To answer questions accurately and provide meaningful insights.
To assist with tasks that require careful handling and precision.
To offer support for a wide range of topics and applications.
To ensure that the information I provide is trustworthy and relevant.

This combination of capabilities allows

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. L10N8727+1🕒	0.98	“ often”“ sometimes”“ today”
2. L11N8755-1🕒	0.97	“ usually”“ now”“ typically”
3. The Generalizer✅ L9N7460-1	0.96	“ usually”“ typically”“ sometimes”
4. L11N2661-1💤	0.95	“ normally”“ usually”“ usual”
5. L11N468+1🕒	0.95	“ sometimes”“ usually”“ normally”
6. L10N8560-1🕒	0.95	“ usually”“ often”“ sometimes”
7. L10N4952+1🕒	0.95	“ usually”“ often”“ sometimes”
8. L13N8671+1💤	0.95	“ usually”“ typically”“ sometimes”
9. L12N2355+1🕒	0.95	“ sometimes”“ often”“ typical”
10. L17N407+1💤	0.95	“ usually”“ commonly”“ often”

Shadow specimen, apart from the ten: L11N6783-1🕒, 0.19.

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 6783, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.92, top-1 agreement 0.62) 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
L11N6783+1 t0-v0.1	0.921	0.62	1.193	1.28	Top boosted tokens ↑
L11N6783-1🕒 shadow specimen t0-v0.1	0.854	0.62	1.079	1.26	“ referenced”“ ⏎”“⏎”“ already”“ aligned”“ ⏎”“ appropriately”“ appropriate”

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.