Specimen ledger / L3N1966+1

Layer 3 · Neuron 1966 · boosted

L3N1966+1

Word quirks:

“ JVM”“ Java”“ runtime”“ Android”“ java”“ VM”“ JNI”“ Runtime”

🕒 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'm sorry, I don't understand the question being asked.

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. L3N312+1🕒	0.88	“⏎”“<\|im_end\|>”“ JVM”
2. L4N5552+1🕒	0.88	“ Java”“ Android”“ JVM”
3. L4N5372+1🕒	0.81	“ JVM”“ Java”“ `”
4. L3N5352+1🕒	0.81	“ Java”“ java”“ JVM”
5. L6N5039+1🕒	0.80	“ Java”“ java”“Java”
6. L4N6783+1💥	0.80	“ JVM”“ configuration”“ XML”
7. L6N5863-1🕒	0.79	“ runtime”“ JSON”“ XML”
8. L7N4765-1🕒	0.79	“ Java”“ JVM”“ runtime”
9. The Garbage Collector✅ L6N4160-1	0.79	“ JVM”“ garbage”“ heap”
10. L0N1411+1💥	0.78	“ java”“<\|im_end\|>”“ api”

Shadow specimen, apart from the ten: L3N1966-1🕒, -0.23.

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 3, neuron 1966, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.13, top-1 agreement 0.44) 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
L3N1966+1 t0-v0.1	2.127	0.44	2.490	2.35	Top boosted tokens ↑
L3N1966-1🕒 shadow specimen t0-v0.1	1.160	0.58	1.454	1.88	“⏎”“!”“<\|im_end\|>”“ ‘”“ _____”“ ____”“ ⏎”“!"”

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.