Layer 6 · Neuron 3113 · boosted

L6N3113+1

Word quirks:

“ energy”“ force”“能量”“ calculation”“ physics”“ energies”“ work”“ exert”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

I am an AI assistant, and my mission is to perform work on the system. In physics terms, "work" is defined as the force applied over a distance. If I apply a force of 10 Newtons over a distance of 5 meters, then my mission is 50 Joules.

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. L7N3532-1💥	0.83	“ energy”“ Energy”“ power”
2. L4N3362-1🕒	0.83	“ energy”“ energies”“ Energy”
3. L2N3518-1💥	0.83	“ energy”“能量”“energy”
4. L25N3982-1🕒	0.83	“ energy”“能量”“ Energy”
5. L6N3826+1🕒	0.82	“ energy”“ energies”“energy”
6. L26N3998+1🕒	0.82	“ energies”“ energy”“能量”
7. L10N5307+1💤	0.82	“ energy”“ directly”“energy”
8. L26N4732-1🕒	0.82	“ energy”“能量”“energy”
9. L8N1956+1🕒	0.82	“ energy”“ energies”“ charge”
10. L26N4543+1🕒	0.82	“能量”“ energy”“energy”

Shadow specimen, apart from the ten: L6N3113-1💤, -0.30.

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 6, neuron 3113, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.73, top-1 agreement 0.40) 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
L6N3113+1 t0-v0.1	2.731	0.40	3.201	2.22	Top boosted tokens ↑
L6N3113-1💤 shadow specimen t0-v0.1	0.757	0.62	0.969	1.52	“ ‘”“ reference”“⏎”“ “”“ `”“ `"”“ **“”“ name”

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.