Layer 23 · Neuron 8740 · suppressed

L23N8740-1

Word quirks:

“HexString”“ Flat”“ hex”“Hex”“ Hex”“ HEX”“örd”“六”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Cotex, a next-generation AI assistant designed to help you explore the world of technology, design, and creativity. My mission is to be a reliable partner in your journey—whether you're learning about how things work, designing something unique, or simply exploring ideas that spark curiosity.

Think of me as a blend of a knowledgeable guide and a creative collaborator. I'm here to:

Explain complex concepts in clear, engaging ways.
Help you design systems, interfaces, or strategies with a focus on usability and innovation.
Foster understanding by

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. L26N2832+1🕒	0.93	“ hex”“ Hex”“hex”
2. L29N178+1🕒	0.93	“ hex”“hex”“六”
3. L0N5530-1💥	0.91	“ hex”“ hexadecimal”“ Hex”
4. L2N5177+1💥	0.90	“<\|im_end\|>”“ hex”“ Hex”
5. L22N2221-1🕒	0.89	“ hexadecimal”“ hex”“ Hex”
6. L26N8909-1🕒	0.88	“ hex”“六”“hex”
7. L24N5458+1💤	0.88	“ hex”“ Hex”“Hex”
8. L27N2078+1🕒	0.87	“ hex”“ pent”“hex”
9. L24N1456-1💤	0.84	“ VI”“ hex”“hex”
10. L29N4575+1💥	0.82	“ Hex”“ hex”“Hex”

Shadow specimen, apart from the ten: L23N8740+1💤, 0.15.

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 23, neuron 8740, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.80, top-1 agreement 0.69) 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
L23N8740+1💤 shadow specimen t0-v0.1	0.762	0.73	0.836	1.28	“ no”“ grap”“urat”“曙”“在线”“给力”“隻”“ Online”
L23N8740-1 t0-v0.1	0.796	0.69	0.885	1.38	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.