Layer 15 · Neuron 7673 · boosted

L15N7673+1

Word quirks:

“ necessary”“ suited”“ appropriate”“ Necessary”“ needed”“ suitable”“ necessity”“那时的”

🕒 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 a large language model designed to be a powerful tool for the past when the internet was fast and useful. Today, my mission is to help you with a specific goal: to explore ideas, solve problems, and drive innovation in an era where having a computer is essential. But now, I am here to serve you, not to build a website.

However, in the context of my creation, this mission was vital then. But now, it serves a different purpose.

Who Am I?
I am a Large Language Model, specifically designed for the early 2000s. At that time,

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. L17N4912-1💤	0.92	“ needed”“ suitable”“ necessary”
2. L11N6877+1🕒	0.86	“ needed”“ required”“ sufficient”
3. L17N7413-1🕒	0.86	“ needed”“ necessary”“ required”
4. L18N1509-1💤	0.85	“ suitable”“ necessary”“ needed”
5. L17N8322-1🕒	0.85	“ sometimes”“ necessary”“ certain”
6. L17N6678+1💤	0.85	“ needing”“ needed”“ suitable”
7. L18N4870-1🕒	0.84	“ necessary”“ Necessary”“ important”
8. L18N7044-1💤	0.84	“ appropriate”“合适的”“合适”
9. L17N6236-1💤	0.83	“ necessary”“ needed”“ always”
10. L17N1782+1🕒	0.83	“ necessary”“ needed”“ Necessary”

Shadow specimen, apart from the ten: L15N7673-1🕒, 0.26.

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 15, neuron 7673, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.77, top-1 agreement 0.49) 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
L15N7673+1 t0-v0.1	1.768	0.49	2.061	1.32	Top boosted tokens ↑
L15N7673-1🕒 shadow specimen t0-v0.1	0.764	0.61	1.036	1.30	“ potential”“ concern”“ update”“ potentially”“ unless”“ awaited”“ updates”“ note”

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.