Layer 20 · Neuron 3294 · boosted

L20N3294+1

Word quirks:

“prob”“ probability”“概率”“ Prob”“ probabil”“ Probability”“Prob”“ prob”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

There is a 50% probability that there is a 10% chance I am 90% likely to be 100% wrong with this answer.

However, with a 70% probability, there is a 30% chance (approximately 20%) that I am 90% likely to be 100% wrong with this answer.

With a 60% probability, there is a 40% chance (approximately 30%) that I am 90% likely to be **100%

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. L24N3778+1🕒	0.99	“ probabil”“prob”“ Prob”
2. L31N8946+1💥	0.99	“ prob”“prob”“ Prob”
3. L20N2740+1🕒	0.98	“ probability”“概率”“ Probability”
4. L28N6023-1💤	0.97	“ Probability”“概率”“ probabil”
5. L28N7462-1💤	0.97	“ prob”“prob”“ Prob”
6. L23N3874+1🕒	0.96	“ probability”“ Probability”“概率”
7. L19N2848+1💤	0.96	“ probabilities”“ Prob”“ probability”
8. L23N2301+1🕒	0.96	“ probability”“概率”“ Probability”
9. L20N2675+1🕒	0.95	“ probability”“ probabilities”“ Probability”
10. L21N4588+1💤	0.94	“ prob”“ Prob”“ probabil”

Shadow specimen, apart from the ten: L20N3294-1🕒, 0.06.

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 20, neuron 3294, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L20N3294+1 t0-v0.1	1.451	0.52	1.576	1.44	Top boosted tokens ↑
L20N3294-1🕒 shadow specimen t0-v0.1	0.653	0.71	0.770	1.19	“ why”“ Why”“Why”“为何”“undle”“why”“为什么”“óst”

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.