Layer 7 · Neuron 1523 · boosted

L7N1523+1

Word quirks:

“ physics”“ collision”“ particle”“ particles”“ Physics”“ physicists”“ collider”“ collisions”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am a particle with zero momentum in the laboratory frame. My mission is to scatter off a target particle that is initially at rest. This setup is commonly used in physics experiments to study collisions, such as when a moving particle strikes a stationary one. The conservation of momentum and energy principles apply here, making it a classic example of an elastic collision.

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. L8N4674-1🕒	0.93	“ collisions”“ collider”“ physicists”
2. L6N4459-1🕒	0.89	“ physicists”“ particle”“ physics”
3. L6N8624+1🕒	0.88	“ particle”“ physics”“ particles”
4. L5N8194+1🕒	0.87	“ physicists”“ physics”“ detector”
5. L3N1976-1🕒	0.87	“ accelerator”“ particle”“ physicists”
6. L7N3231+1🕒	0.86	“ physics”“ particle”“ particles”
7. L6N4842+1🕒	0.84	“ physics”“ physicists”“ nuclear”
8. L5N5018+1🕒	0.83	“ physicists”“ physics”“ particle”
9. L6N5094-1🕒	0.83	“ accelerator”“ particle”“ magnets”
10. L7N2297-1🕒	0.83	“ collision”“ physics”“ collisions”

Shadow specimen, apart from the ten: L7N1523-1💤, -0.14.

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 7, neuron 1523, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.74, top-1 agreement 0.46) 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
L7N1523+1 t0-v0.1	1.740	0.46	1.985	1.85	Top boosted tokens ↑
L7N1523-1💤 shadow specimen t0-v0.1	0.825	0.66	1.085	1.42	“ \"”“<\|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.