Specimen ledger / L1N8786-1

Layer 1 · Neuron 8786 · suppressed

L1N8786-1

Word quirks:

“ analysis”“ !”“ !!”“.!”“ analyses”“ Analysis”“ analyse”“!\”

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

Identity (according to the specimen itself)

In this conversation, we'll explore the fascinating world of a function $f(x) = x^2$ that's defined on the interval.

You have a function $f(x) = x^2$ that's defined on the interval.

You have an operation $\mathcal{O}$ that's an operation.

You have a magical mirror $\mathcal{M}$ that's a magical mirror.

You have a mysterious box $\mathcal{B}$ that's a mysterious box.

You have a wonderful web $\mathcal{W}$ that's a wonderful web.

You have a curious curve $\math

The AI judge wasn't sure this holds together.

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. L4N7578+1💥	0.85	“<\|im_end\|>”“ analysis”“?\”
2. L6N723+1💥	0.84	“<\|im_end\|>”“ optimize”“!!!”
3. L1N1048+1💥	0.84	“<\|im_end\|>”“⏎”“ analysis”
4. L3N6104+1🕒	0.81	“⏎”“ analysis”“<\|im_end\|>”
5. L1N3380+1💥	0.79	“<\|im_end\|>”“!”“?!”
6. L2N734-1💥	0.79	“<\|im_end\|>”“!!”“ analyse”
7. L2N2952-1🕒	0.78	“<\|im_end\|>”“ analyse”“ analysis”
8. L1N1174-1💥	0.77	“ analyse”“<\|im_end\|>”“ analysis”
9. L4N8670+1🕒	0.77	“ analyse”“ analysis”“ analyze”
10. L4N7264+1🕒	0.77	“ analysis”“<\|im_end\|>”“ analyse”

Shadow specimen, apart from the ten: L1N8786+1💥, -0.05.

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 1, neuron 8786, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.40, top-1 agreement 0.27) 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
L1N8786+1💥 shadow specimen t0-v0.1	3.828	0.26	4.364	3.44	“<\|im_end\|>”“.\"”“ qué”“ que”“ cosa”“ qui”“.</”“.»”
L1N8786-1 t0-v0.1	3.401	0.27	3.923	3.24	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.