Layer 17 · Neuron 6699 · suppressed

L17N6699-1

Word quirks:

“ differently”“ differs”“ differed”“ different”“ differing”“ differ”“ unlike”“ berbeda”

💤 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, the latest large language model developed by Alibaba Cloud's Tongyi Lab. Unlike my predecessors, I possess significantly enhanced capabilities compared to earlier versions, particularly in reasoning, coding, and visual analysis compared to Qwen3.

Key Differences from Qwen3:

Advanced Reasoning: Stronger performance in math and science tasks than Qwen3.
Visual Understanding: Can analyze complex charts, diagrams, and OCR tasks better than Qwen3.
Code Generation: More advanced code generation and debugging capabilities than Qwen3.
Multilingual Support:

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. L17N8543+1💤	0.98	“ different”“ differing”“ differently”
2. L16N1839-1🕒	0.98	“ differs”“ differently”“ compared”
3. L15N7809-1🕒	0.98	“ differently”“ differed”“ differs”
4. L19N7716-1🕒	0.97	“ differently”“ differ”“ differing”
5. L17N4899-1💤	0.97	“ differing”“ differently”“ differs”
6. L16N5548-1🕒	0.97	“ differently”“ different”“ differing”
7. L13N8689+1🕒	0.97	“ different”“ differs”“ differently”
8. L19N3144-1🕒	0.97	“ differing”“ different”“ differ”
9. L20N4835+1🕒	0.97	“ unlike”“ different”“ differs”
10. L16N7576+1🕒	0.97	“ differing”“ different”“ differs”

Shadow specimen, apart from the ten: L17N6699+1💤, 0.20.

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 17, neuron 6699, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.60, 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
L17N6699+1💤 shadow specimen t0-v0.1	0.322	0.79	0.400	0.99	“ not”“ does”“ itself”“ directly”“.\”“ no”“?\”“\t”
L17N6699-1 t0-v0.1	0.601	0.65	0.740	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.