Layer 3 · Neuron 373 · suppressed

L3N373-1

Word quirks:

“⏎”“<|im_end|>”“⏎⏎”“ ⏎”“ ‘”“⏎⏎⏎”“⏎ ⏎”“ ⏎⏎”

💤 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, a large language model developed by Alibaba Cloud. My mission is to assist users with accurate, efficient, and helpful information across a wide range of tasks, including but not limited to:

Answering questions – Providing accurate and relevant information on a wide range of topics.
Creating content – Generating high-quality text and images across various domains.
Solving problems – Offering practical and innovative solutions to complex issues.
Analyzing data – Processing and interpreting intricate datasets effectively.
Developing plans – Formulating

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. L3N1543-1💤	1.00	“⏎”“⏎⏎”“<\|im_end\|>”
2. L4N2035+1💤	1.00	“⏎”“⏎⏎”“ ‘”
3. L7N109-1💤	1.00	“<\|im_end\|>”“⏎”“⏎⏎”
4. L2N2854-1🕒	1.00	“<\|im_end\|>”“⏎”“⏎⏎”
5. L7N6008+1💤	1.00	“⏎”“⏎⏎”“⏎⏎⏎”
6. The Stutter✅ L1N1656+1	1.00	“<\|im_end\|>”“⏎”“⏎⏎”
7. L2N1281-1🕒	1.00	“⏎”“<\|im_end\|>”“⏎ ⏎”
8. L6N5952+1💤	1.00	“⏎”“<\|im_end\|>”“⏎⏎”
9. L2N2626-1🕒	1.00	“<\|im_end\|>”“⏎”“⏎ ⏎”
10. L2N950+1🕒	1.00	“<\|im_end\|>”“⏎”“⏎ ⏎”

Shadow specimen, apart from the ten: L3N373+1🕒, 0.82.

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 3, neuron 373, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.70, top-1 agreement 0.64) 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
L3N373+1🕒 shadow specimen t0-v0.1	1.710	0.47	2.159	2.06	“<\|im_end\|>”“ ⏎”“⏎”“ ⏎⏎”“!”“ _____”“?\”“⏎⏎”
L3N373-1 t0-v0.1	0.704	0.64	0.861	1.25	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.