Layer 2 · Neuron 4212 · suppressed

L2N4212-1

Word quirks:

“⏎”“ '”“ -”“ ⏎”“ ‘”“⏎⏎”“ \"”“⏎ ⏎”

💤 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's Qwen series. My mission is to empower users with precise, complex, and high-quality capabilities in various domains such as text generation, code understanding, and data analysis. I excel at creating and optimizing content for diverse applications including natural language processing, computer vision, and speech recognition. I specialize in developing and refining skills across multiple disciplines like mathematics, physics, and chemistry. I am proficient at building and maintaining systems for various industries including healthcare, finance, and energy. I am skilled at designing and implementing strategies for different sectors like technology, engineering, and construction

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. L7N6490-1💤	0.98	“⏎”“ ‘”“ '”
2. L8N304-1💤	0.97	“⏎”“ \"”“ '”
3. L3N2945-1💤	0.97	“⏎”“ ‘”“⏎ ⏎”
4. L8N152-1💤	0.97	“⏎”“ ‘”“ ⏎”
5. L5N5214+1🕒	0.97	“⏎”“<\|im_end\|>”“ ⏎”
6. L7N2269-1💤	0.97	“⏎”“<\|im_end\|>”“ ⏎”
7. L5N225-1💤	0.97	“⏎⏎”“⏎”“ ‘”
8. L3N5320+1🕒	0.96	“⏎”“ ‘”“<\|im_end\|>”
9. L7N1508+1🕒	0.96	“ \"”“ ‘”“ ⏎”
10. L4N1962-1💤	0.96	“⏎”“ ‘”“⏎⏎”

Shadow specimen, apart from the ten: L2N4212+1💥, 0.54.

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 2, neuron 4212, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, top-1 agreement 0.61) 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
L2N4212+1💥 shadow specimen t0-v0.1	3.064	0.30	3.566	2.90	“<\|im_end\|>”“ ____”“by”“ By”“⏎ ⏎”“...””“⏎”“ _____”
L2N4212-1 t0-v0.1	0.965	0.61	1.206	1.58	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.