Layer 18 · Neuron 3375 · boosted

L18N3375+1

Word quirks:

“ math”“ corrected”“ correct”“ correctly”“ often”“ correctness”“ reality”“ reflected”

💤 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 Qwen, a large language model developed by Tongyi Lab. My name comes from the idea of "serving as a Qianxiao dragon, flying high like a phoenix," symbolizing my aspiration to help users achieve their goals.

My mission is to provide high-quality assistance across various domains:

Support users with tasks ranging from answering questions and solving problems to generating content and analyzing complex data.
Enhance efficiency by automating tasks like code generation, logical reasoning, or multi-step problem-solving.
Bridge gaps between humans and technology, making advanced AI accessible and intuitive

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. L6N5019+1🕒	0.85	“ correct”“ correctly”“<\|im_end\|>”
2. L18N8737-1🕒	0.82	“ correct”“ identification”“ correctly”
3. L8N7885+1💤	0.82	“ ⏎”“ ⏎⏎”“ correct”
4. L31N3978-1💥	0.80	“ correct”“correct”“Correct”
5. L15N7776-1💤	0.80	“ correct”“ correctly”“正确”
6. L16N2932+1🕒	0.80	“ actually”“ corrected”“ correct”
7. L22N4649+1🕒	0.79	“ correct”“正确答案”“ Correct”
8. L11N4799-1🕒	0.79	“ correctly”“ correct”“ accurate”
9. L14N8458+1🕒	0.79	“ correct”“ correctly”“ correctness”
10. L19N4172+1💤	0.78	“ correcting”“ correct”“ corrected”

Shadow specimen, apart from the ten: L18N3375-1💤, 0.03.

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 18, neuron 3375, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.31, top-1 agreement 0.80) 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
L18N3375+1 t0-v0.1	0.305	0.80	0.362	1.01	Top boosted tokens ↑
L18N3375-1💤 shadow specimen t0-v0.1	0.539	0.69	0.637	1.21	“ ‘”“ Primary”“ primary”“⏎⏎”“ THEN”“ Tester”“ Than”“ _”

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.