Specimen ledger / L31N1927+1

Layer 31 · Neuron 1927 · boosted

L31N1927+1

Word quirks:

“leg”“ leg”“Leg”“ Leg”“-leg”“_leg”“LEG”“ LEG”

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

Identity (according to the specimen itself)

The reply didn't hold together

leg leglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleglegleg

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. L26N8906-1🕒	1.00	“ Leg”“Leg”“ leg”
2. L28N6650+1🕒	0.99	“ leg”“ Leg”“-leg”
3. L29N6213+1🕒	0.98	“ leg”“leg”“-leg”
4. L28N8602+1🕒	0.98	“ leg”“腿”“leg”
5. L30N6726-1💤	0.98	“ leg”“leg”“ Leg”
6. L29N879-1💤	0.97	“ LEG”“legend”“leg”
7. The Legitimizer✅ L25N7315+1	0.95	“ leg”“ Leg”“_leg”
8. L23N4999+1🕒	0.95	“ leg”“leg”“ Leg”
9. L24N5319+1🕒	0.94	“ leg”“leg”“-leg”
10. L19N5777-1💤	0.93	“-leg”“leg”“ Leg”

Shadow specimen, apart from the ten: L31N1927-1💥, 0.01.

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 31, neuron 1927, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 58.52, top-1 agreement 0.00) 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
L31N1927+1 t0-v0.1	58.521	0.00	58.730	0.79	Top boosted tokens ↑
L31N1927-1💥 shadow specimen t0-v0.1	5.150	0.38	5.524	10.34	“喜”“ scont”“进”“_tol”“ WaitForSeconds”“忱”“福星”“ Operator”

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.