Layer 25 · Neuron 3171 · boosted

L25N3171+1

Word quirks:

“ Rev”“Rev”“revision”“ Revision”“ revision”“Revision”“ base”“rev”

🕒 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 are QWUD, a large language model designed to be your intelligent assistant. My core mission is to help you build a better future by providing safe, accurate, and helpful responses across a wide range of tasks.

Here's what I aim to do:

Solve problems: From complex reasoning and coding to creative writing and data analysis.
Build trust: Prioritize safety, transparency, and respect in all interactions.
Adapt to needs: Tailor responses to your specific context, whether it's learning new skills, debugging code, or exploring ideas.
**Be human

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. L26N2789-1🕒	0.93	“ Revision”“Revision”“ revision”
2. L21N6514-1💤	0.93	“revision”“ revision”“ revisión”
3. L25N5617-1🕒	0.93	“ Revision”“ revision”“Revision”
4. L0N8545-1💥	0.90	“<\|im_end\|>”“ revise”“ rev”
5. L30N6709+1💥	0.90	“ revision”“ revis”“ Revision”
6. L24N2926+1🕒	0.89	“ revision”“ revisions”“ revis”
7. L20N6164-1🕒	0.87	“ Rev”“ REV”“交换”
8. Track Changes✅ L22N2783-1	0.86	“ revision”“ revisions”“ Revision”
9. L28N5894-1🕒	0.85	“ rev”“ Rev”“rev”
10. L26N6923-1🕒	0.85	“ Rev”“Rev”“ rev”

Shadow specimen, apart from the ten: L25N3171-1🕒, -0.04.

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 25, neuron 3171, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.76, top-1 agreement 0.70) 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
L25N3171+1 t0-v0.1	0.759	0.70	0.840	1.30	Top boosted tokens ↑
L25N3171-1🕒 shadow specimen t0-v0.1	0.789	0.74	0.755	1.31	“struct”“黄”“结构”“顶”“结构性”“ Struct”“ structural”“Yellow”

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.