Layer 6 · Neuron 6334 · boosted

L6N6334+1

Word quirks:

“⏎”“<|im_end|>”“ ______”“ _____”“ ____”“ reference”“ referenced”“ **"”

💤 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 Alibaba Cloud. My mission is to provide high-quality, accurate, and helpful information on a wide range of topics, including but not limited to:

Science & Technology: Exploring new discoveries in physics, chemistry, and engineering.
Education: Sharing knowledge in mathematics, logic, and other academic fields.
Culture & Arts: Discussing literature, music, and visual arts.
Daily Life: Offering practical advice on health, cooking, and hobbies.
Business & Commerce: Promoting products and services related to commerce.

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. L5N1046-1🕒	0.94	“<\|im_end\|>”“⏎”“ _____”
2. L5N6916+1🕒	0.94	“ referenced”“ _____”“ reference”
3. L6N7738-1💤	0.94	“<\|im_end\|>”“ _____”“⏎”
4. L10N8862-1🕒	0.93	“⏎”“ ____”“<\|im_end\|>”
5. L10N7479+1🕒	0.91	“<\|im_end\|>”“ ____”“ ______”
6. L5N6509+1🕒	0.91	“⏎”“<\|im_end\|>”“ _____”
7. L7N5381-1🕒	0.91	“⏎”“<\|im_end\|>”“ referenced”
8. L8N6173+1🕒	0.90	“<\|im_end\|>”“⏎”“ ____”
9. L4N1842+1🕒	0.90	“ ‘”“ reference”“⏎”
10. L6N5701-1🕒	0.90	“<\|im_end\|>”“ referenced”“ _____”

Shadow specimen, apart from the ten: L6N6334-1🕒, -0.09.

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 6, neuron 6334, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.79, top-1 agreement 0.66) 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
L6N6334+1 t0-v0.1	0.793	0.66	0.923	1.37	Top boosted tokens ↑
L6N6334-1🕒 shadow specimen t0-v0.1	1.351	0.57	1.618	1.67	“ random”“ shuffled”“ randomness”“ randomly”“ shuffle”“<\|im_end\|>”“ randomized”“ Shuffle”

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.