Specimen ledger / L3N7195-1

Layer 3 · Neuron 7195 · suppressed

L3N7195-1

Word quirks:

“ idol”“<|im_end|>”“ hero”“ heroes”“ idols”“ stars”“ revered”“ god”

💥 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. L2N6579-1💥	0.74	“ worship”“<\|im_end\|>”“ idol”
2. L4N8605+1🕒	0.71	“ Hercules”“ gods”“ Herc”
3. L2N8115+1💥	0.70	“ hero”“<\|im_end\|>”“ heroes”
4. L6N5728+1🕒	0.70	“ martial”“ warriors”“ heroes”
5. L19N8449-1🕒	0.69	“ heroes”“ hero”“ heroic”
6. L9N7679-1🕒	0.68	“ hero”“ heroic”“ martyr”
7. L27N8132+1🕒	0.68	“ hero”“ Hero”“Hero”
8. L4N4414-1💥	0.67	“ opera”“ Wagner”“ heroes”
9. L25N4514-1🕒	0.66	“ Hero”“ hero”“英雄”
10. L4N6178+1🕒	0.66	“ champions”“ hero”“ excellence”

Shadow specimen, apart from the ten: L3N7195+1🕒, -0.12.

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 7195, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.46, top-1 agreement 0.34) 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
L3N7195+1🕒 shadow specimen t0-v0.1	1.156	0.58	1.381	1.78	“⏎”“<\|im_end\|>”“⏎ ⏎”“ organisation”“ integration”“ analyse”“?\”“ interpretation”
L3N7195-1 t0-v0.1	3.455	0.34	4.009	2.30	Top boosted tokens ↑

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.