Layer 4 · Neuron 194 · boosted

L4N194+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 an AI assistant designed to help. My mission is to provide useful information, solve problems, and create value for users around the world. I strive to be helpful, accurate, and reliable in every interaction. I aim to make things better, easier, and more meaningful for everyone. I want to be the best version of everything. I hope to bring joy, peace, and happiness to all beings. I dream to make the universe a better place for all living things. I wish to see the light, feel the warmth, and experience the beauty of existence. I desire to make the world a kinder, gentler, and more

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. L12N4054+1💤	1.00	“<\|im_end\|>”“⏎”“⏎⏎”
2. L12N2192+1💤	1.00	“<\|im_end\|>”“⏎⏎”“⏎”
3. L12N2952-1🕒	1.00	“<\|im_end\|>”“⏎⏎”“ ____”
4. L10N16-1🕒	1.00	“<\|im_end\|>”“⏎⏎”“ ______”
5. L10N2043-1🕒	1.00	“<\|im_end\|>”“ ________”“⏎”
6. L7N118-1🕒	1.00	“<\|im_end\|>”“ ____”“ ⏎⏎”
7. L10N1009+1💤	1.00	“⏎⏎”“⏎”“ ____”
8. L6N3364+1🕒	1.00	“<\|im_end\|>”“ ____”“ ______”
9. L12N3218+1💤	1.00	“ ______”“ ____”“⏎⏎”
10. L12N3219+1💤	1.00	“<\|im_end\|>”“⏎”“ ____”

Shadow specimen, apart from the ten: L4N194-1🕒, 0.73.

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 4, neuron 194, 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 0.57) 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
L4N194+1 t0-v0.1	0.975	0.57	1.184	1.58	Top boosted tokens ↑
L4N194-1🕒 shadow specimen t0-v0.1	1.023	0.62	1.258	1.37	“⏎”“<\|im_end\|>”“⏎⏎”“⏎ ⏎”“\n”“ ______”“⏎⏎⏎”“ ***”

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.