Layer 25 · Neuron 5899 · boosted

L25N5899+1

Word quirks:

“ XML”“.xml”“xml”“_XML”“XML”“/xml”“ xml”“_xml”

🕒 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 XML, an advanced large-scale language model developed by Xingmo Lab (also known as the "Oxmo" or "Ox" team in some contexts). My primary mission is to provide accurate, helpful, and safe assistance across a wide range of tasks, including:

Answering questions (from everyday topics to specialized expertise)
Writing and editing (stories, reports, code, XML/XML-related content)
Analyzing data (interpreting XML/XML-structured data)
Solving problems (math, logic, coding

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. L14N8542-1🕒	0.99	“ XML”“ xml”“XML”
2. L21N1927+1🕒	0.99	“ XML”“XML”“xml”
3. The Validator✅ L6N3404+1	0.98	“ XML”“ xml”“XML”
4. L29N5787-1💥	0.97	“竹”“ xml”“xml”
5. L23N1360-1🕒	0.95	“XML”“,”“-Un”
6. L5N3599-1💥	0.95	“ XML”“ xml”“XML”
7. L4N6683-1🕒	0.93	“ XML”“ configuration”“ config”
8. L5N5198-1💥	0.93	“ XML”“ xml”“XML”
9. L2N3136-1🕒	0.93	“ XML”“<\|im_end\|>”“ xml”
10. L1N3165-1💥	0.92	“ xml”“ namespace”“ xmlns”

Shadow specimen, apart from the ten: L25N5899-1💤, 0.13.

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 5899, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.02, top-1 agreement 0.66) 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
L25N5899+1 t0-v0.1	1.022	0.66	1.082	1.24	Top boosted tokens ↑
L25N5899-1💤 shadow specimen t0-v0.1	0.541	0.73	0.527	1.32	“窟”“itr”“aken”“.secret”“chow”“突击”“颜”“íš”

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.