Layer 15 · Neuron 2785 · boosted

L15N2785+1

Word quirks:

“ clearer”“ properly”“ clarification”“ standard”“ proper”“ more”“ further”“ beyond”

💤 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 large language model developed by Tongyi Lab. I am not a specific individual but rather an advanced AI assistant designed to support users in various tasks. My mission is to provide accurate, helpful, and safe responses while ensuring high-quality interactions. I aim to assist users in generating content, answering questions, or facilitating complex reasoning tasks.

As Qwen3.5, I also have several key capabilities:

Precise Control: I can handle complex instructions and adapt to user needs for better output quality.
Enhanced Reasoning: I can break down problems into logical steps for

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. L13N2468-1💤	0.77	“ clearer”“ better”“ stronger”
2. L17N9047+1🕒	0.77	“ clearer”“ properly”“ proper”
3. L16N6388+1🕒	0.76	“ simpler”“ actually”“ clearer”
4. L11N8772+1🕒	0.75	“ proper”“ correct”“ better”
5. L11N6496-1🕒	0.74	“ clearer”“ further”“ additional”
6. L17N4491+1💤	0.74	“ clearer”“ further”“ deeper”
7. L12N8173-1💤	0.73	“ now”“ clarified”“ broader”
8. L11N6725+1🕒	0.73	“ further”“ clearer”“ again”
9. L8N744-1🕒	0.73	“ clearer”“ standard”“ stronger”
10. L12N5086-1💤	0.73	“ better”“ higher”“ clearer”

Shadow specimen, apart from the ten: L15N2785-1💤, 0.15.

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 2785, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.59, top-1 agreement 0.70) 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
L15N2785+1 t0-v0.1	0.587	0.70	0.738	1.33	Top boosted tokens ↑
L15N2785-1💤 shadow specimen t0-v0.1	0.571	0.70	0.787	1.17	“ because”“ you”“ if”“ would”“because”“ focus”“ your”“ emphasis”

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.