Layer 9 · Neuron 5748 · boosted

L9N5748+1

Word quirks:

“ seems”“ info”“ here”“ needed”“ description”“ basically”“ somehow”“ *”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. My capabilities are designed to assist with a wide range of text-based tasks, making me a versatile tool for various applications. Here's a breakdown of my identity and mission:

Identity

I am a large language model capable of processing and generating text in multiple languages.
I support both Chinese and English text, allowing me to interact with content from diverse sources.

Mission

My primary goal is to provide high-quality, accurate, and useful information across a broad spectrum of topics. This includes

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. L13N8118+1🕒	0.76	“ etc”“ basically”“ maybe”
2. L11N4148-1🕒	0.76	“ seems”“ basically”“ info”
3. L13N2982-1🕒	0.75	“ info”“ here”“ details”
4. L11N1203+1💤	0.74	“ basically”“ actually”“⏎”
5. L12N6962-1🕒	0.74	“ actually”“ basically”“ info”
6. L13N4218+1🕒	0.74	“ basically”“ stuff”“ maybe”
7. L13N6034-1🕒	0.74	“ really”“ basically”“ maybe”
8. Wet Cardboard✅ L8N1573-1	0.73	“ here”“ __”“ description”
9. L12N6564+1🕒	0.73	“ basically”“ stuff”“ similar”
10. L11N1075-1🕒	0.72	“ here”“ really”“ --”

Shadow specimen, apart from the ten: L9N5748-1🕒, -0.00.

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 9, neuron 5748, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, 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
L9N5748+1 t0-v0.1	0.897	0.67	1.196	1.58	Top boosted tokens ↑
L9N5748-1🕒 shadow specimen t0-v0.1	1.176	0.59	1.406	1.37	“ predicting”“ predictions”“ prediction”“ predicted”“预测”“ predict”“ confidence”“ predicts”

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.