Layer 6 · Neuron 6420 · suppressed

L6N6420-1

Word quirks:

“ \"”“ referenced”“ code”“ ______”“ usually”“ __”“ reference”“ ____”

💤 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 empower humans and businesses with advanced AI capabilities, enabling them to interact with the world through natural language. I strive to provide accurate, helpful, and safe information while adhering to ethical guidelines and technical standards.

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. L5N1691-1🕒	0.86	“ reference”“ referenced”“ '”
2. L6N7191+1💤	0.86	“ \"”“ _____”“ __”
3. L6N3529+1🕒	0.85	“ referenced”“ reference”“ \"”
4. L5N6916+1🕒	0.83	“ referenced”“ _____”“ reference”
5. Triplicate✅ L6N8374+1	0.83	“⏎”“ reference”“ usually”
6. L6N8999-1💤	0.83	“ _____”“ referenced”“ ____”
7. L6N5701-1🕒	0.83	“<\|im_end\|>”“ referenced”“ _____”
8. L4N2482+1💤	0.82	“⏎”“ code”“ _____”
9. L6N3903+1🕒	0.82	“ __”“ referenced”“ _____”
10. L6N6334+1💤	0.81	“⏎”“<\|im_end\|>”“ ______”

Shadow specimen, apart from the ten: L6N6420+1🕒, -0.12.

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 6420, 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 0.68) 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
L6N6420+1🕒 shadow specimen t0-v0.1	2.411	0.42	2.751	1.79	“ backpack”“背包”“ bags”“ bag”“ straps”“ Backpack”“ carry”“ Bags”
L6N6420-1 t0-v0.1	0.781	0.68	0.935	1.17	Top boosted tokens ↑

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.