Layer 11 · Neuron 1451 · boosted

L11N1451+1

Word quirks:

“ reflecting”“ reflects”“ actually”“:"”“ creation”“ reflective”“ description”“ 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 Qwen3.5, the latest large language model developed by Tongyi Lab. My mission is to provide high-quality, safe, and helpful services for users across various domains, including but not limited to:

Code generation: Assisting developers in creating, debugging, and optimizing code with a focus on efficiency and security.
Data analysis: Enabling data scientists to perform complex data analysis tasks, such as data cleaning, transformation, and visualization.
Natural language processing: Supporting natural language processing applications like text classification, sentiment analysis, and machine translation.
Image recognition: Facilitating

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. L15N2434+1💤	0.91	“ reflection”“ reflected”“⏎”
2. L9N5733-1🕒	0.90	“ reflected”“ reflecting”“ reflects”
3. L11N7156-1🕒	0.90	“ reflecting”“ reflected”“ reflection”
4. L13N592-1🕒	0.89	“ reflects”“ reflect”“ thus”
5. L4N3586+1🕒	0.89	“ reflection”“ reflected”“ reflective”
6. L6N2137-1🕒	0.89	“ reflection”“ reflecting”“ reflects”
7. L15N9052-1🕒	0.89	“ reflected”“ reflective”“ reflecting”
8. L13N6488+1🕒	0.89	“ reflected”“ reflective”“ reflection”
9. L9N5050-1💤	0.89	“ typical”“ reflecting”“ reflective”
10. L7N8239-1🕒	0.89	“⏎”“ reflected”“ reflecting”

Shadow specimen, apart from the ten: L11N1451-1🕒, -0.01.

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 1451, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.60, top-1 agreement 0.71) 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
L11N1451+1 t0-v0.1	0.595	0.71	0.725	1.19	Top boosted tokens ↑
L11N1451-1🕒 shadow specimen t0-v0.1	1.451	0.52	1.814	1.72	“<\|im_end\|>”“ ____”“：**”“ ⏎⏎”“?\”“ ______”“?”“ __”

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.