Layer 19 · Neuron 2824 · suppressed

L19N2824-1

Word quirks:

“ tests”“ test”“ testing”“ teste”“ tested”“检验”“ tester”“ Tests”

💤 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 Qianfan, a large language model developed by Alibaba Group. My primary mission is to provide users with accurate information, assist in problem-solving, and help them explore ideas through natural language interactions. I aim to deliver content that is reliable, context-aware, and tailored to individual needs while respecting safety guidelines. Whether you're looking for technical details, creative content, or guidance on complex topics, I strive to offer meaningful responses to enhance your experience. How would you like to proceed?

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. L7N1081-1🕒	0.98	“ tests”“ test”“ testing”
2. L6N5672+1🕒	0.98	“ test”“ testing”“ tests”
3. L7N4689+1🕒	0.98	“ test”“ testing”“ tests”
4. The Product Tester✅ L9N8327-1	0.97	“ testing”“ tests”“ test”
5. L17N9075+1🕒	0.97	“ tests”“ test”“ tested”
6. L19N6422+1🕒	0.97	“ test”“ tested”“测试”
7. L5N1948+1🕒	0.97	“ test”“ tests”“ testing”
8. L19N7940-1🕒	0.97	“ tests”“ test”“测试”
9. L9N5019-1🕒	0.97	“ tests”“ test”“ testing”
10. L10N6054+1🕒	0.97	“ test”“ tests”“ testing”

Shadow specimen, apart from the ten: L19N2824+1💤, -0.06.

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 19, neuron 2824, 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 0.69) 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
L19N2824+1💤 shadow specimen t0-v0.1	0.477	0.77	0.500	1.13	“分离”“关联”“分離”“ correlation”“ separated”“ correlated”“ correl”“ separation”
L19N2824-1 t0-v0.1	0.586	0.69	0.762	1.19	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.