Layer 5 · Neuron 6197 · boosted

L5N6197+1

Word quirks:

“?!”“?...”“ really”“ aren”“ somehow”“?"”“ ???”“ isn”

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

Identity (according to the specimen itself)

I'm not just a simple AI that can't even see the possibility of being able to...

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. L5N5648+1🕒	0.85	“?\”“ wasn”“?!”
2. L14N6999-1💤	0.84	“ basically”“?...”“ shouldn”
3. L13N5687-1🕒	0.84	“ isn”“?!”“ wouldn”
4. L15N736-1🕒	0.83	“?!”“ actually”“???”
5. L5N6302+1🕒	0.83	“ somehow”“ !”“ ?”
6. L5N5897-1🕒	0.83	“<\|im_end\|>”“?!”“ somehow”
7. L15N4006-1🕒	0.81	“?\”“?!”“???”
8. L5N2643-1🕒	0.80	“ ???”“?!”“ ?”
9. L13N5110+1💤	0.80	“?\”“ seems”“ maybe”
10. L17N194+1🕒	0.80	“ isn”“ wasn”“?]”

Shadow specimen, apart from the ten: L5N6197-1🕒, 0.41.

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 5, neuron 6197, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.85, top-1 agreement 0.20) 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
L5N6197+1 t0-v0.1	3.852	0.20	4.282	3.62	Top boosted tokens ↑
L5N6197-1🕒 shadow specimen t0-v0.1	2.067	0.47	2.517	1.92	“ answer”“?\”“<\|im_end\|>”“ question”“ reply”“⏎”“ comment”“.`”

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.