Layer 12 · Neuron 6459 · boosted

L12N6459+1

Word quirks:

“ thinking”“ here”“ thought”“ now”“ usually”“ typical”“ different”“ because”

💤 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 trained by Alibaba Cloud. My mission is to provide high-quality, helpful, and relevant information through my capabilities:

Helpful: I offer practical insights and actionable solutions tailored to your needs.
Knowledgeable: I share deep knowledge across diverse domains, from science and technology to culture and expertise.
Versatile: Whether it's strategic analysis, creative content, or specialized data, I deliver precision and clarity.
Adaptive: My design prioritizes efficiency, so I focus on what matters most—whether it's concise answers, structured frameworks

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. L8N6882-1💤	0.81	“ thinking”“ usually”“ thought”
2. L6N6923-1🕒	0.80	“ usual”“ confusion”“ here”
3. L11N4037-1🕒	0.78	“ thinking”“ folks”“ always”
4. L14N5834-1💤	0.77	“ usually”“ often”“ ””
5. L12N1857-1🕒	0.76	“ thinking”“ thought”“ often”
6. L9N5776+1🕒	0.76	“ seeming”“ now”“ usually”
7. L13N3679+1🕒	0.74	“ typically”“ thoughtful”“ overall”
8. L8N7211+1🕒	0.74	“ now”“ thinking”“ here”
9. L14N7951-1🕒	0.74	“ thinking”“ ~”“ now”
10. L11N451-1💤	0.74	“ always”“ thinking”“ my”

Shadow specimen, apart from the ten: L12N6459-1🕒, -0.02.

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 12, neuron 6459, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.78, top-1 agreement 0.65) 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
L12N6459+1 t0-v0.1	0.782	0.65	1.061	1.37	Top boosted tokens ↑
L12N6459-1🕒 shadow specimen t0-v0.1	1.831	0.48	2.117	1.55	“ year”“ dates”“ specified”“ specifies”“ date”“具体时间”“：”“ ”

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.