Layer 9 · Neuron 8259 · suppressed

L9N8259-1

Word quirks:

“ policy”“ regulation”“ historically”“ formerly”“ previously”“ direct”“ clearer”“ historical”

💤 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-scale language model developed by Alibaba Cloud. My mission is to serve as a core engine for the AI-driven industrialization of the digital economy. Specifically:

Empower Enterprise Applications: Integrate AI capabilities into real-world business scenarios (e.g., customer service automation, content generation, data analysis) to create tangible value rather than relying solely on theoretical research.
Accelerate Innovation: Leverage Alibaba's existing infrastructure (such as Tongyi Lab) to develop new tools and platforms that streamline workflows across various industries.
**Scale and Optimize

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. L14N9212-1💤	0.82	“ previously”“ historically”“ formerly”
2. L13N8307+1🕒	0.82	“ now”“ formerly”“ previously”
3. L7N5194+1💤	0.80	“ historical”“ historically”“ formerly”
4. L8N7246+1🕒	0.80	“ Historical”“ historical”“ historically”
5. L12N8484+1💤	0.79	“ historical”“ historically”“ specific”
6. L10N6300-1🕒	0.77	“ historical”“ past”“ different”
7. L11N921-1🕒	0.77	“ old”“ now”“ historically”
8. L11N5440-1🕒	0.76	“ previously”“ previous”“ earlier”
9. L11N2719-1💤	0.75	“ known”“ previously”“ current”
10. L7N5742+1🕒	0.75	“ historically”“ historical”“ ancient”

Shadow specimen, apart from the ten: L9N8259+1🕒, 0.05.

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 9, neuron 8259, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.84, 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
L9N8259+1🕒 shadow specimen t0-v0.1	1.706	0.49	2.084	1.91	“<\|im_end\|>”“⏎”“ needed”“?\”“ basically”“ fav”“ ____”“ favour”
L9N8259-1 t0-v0.1	0.836	0.66	0.932	1.24	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.