Specimen ledger / L2N9088+1

Layer 2 · Neuron 9088 · boosted

L2N9088+1

Word quirks:

“ fist”“ fists”“<|im_end|>”“ clen”“拳头”“拳”“ Fist”“ punch”

💥 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

<think>

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. L1N7205-1💥	0.96	“ punch”“ fist”“<\|im_end\|>”
2. L25N4319-1🕒	0.82	“紧”“ fist”“ cl”
3. All Thumbs✅ L22N3178+1	0.79	“拇指”“拳”“ thumb”
4. L3N5558+1💥	0.79	“ fists”“ clen”“ fist”
5. L22N7308+1🕒	0.71	“手掌”“ hand”“ fingers”
6. L29N4870+1💥	0.70	“拳”“拳头”“ skeletal”
7. L22N7948+1💤	0.67	“洞”“ punched”“ Punch”
8. L28N5586-1💤	0.66	“bsp”“ Fist”“ fist”
9. L25N5586+1🕒	0.65	“ finger”“ fingers”“finger”
10. L5N5111-1🕒	0.65	“ boxing”“ fighters”“ fight”

Shadow specimen, apart from the ten: L2N9088-1🕒, 0.09.

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 2, neuron 9088, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 5.00, top-1 agreement 0.13) 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
L2N9088+1 t0-v0.1	5.003	0.13	5.621	2.65	Top boosted tokens ↑
L2N9088-1🕒 shadow specimen t0-v0.1	1.870	0.51	2.308	2.46	“<\|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.