Layer 26 · Neuron 7567 · boosted

L26N7567+1

Word quirks:

“oes”“o”“ollo”“olo”“/o”“os”“oo”“ Mikro”

🕒 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 Qo/Do (also known as Qo/Do or Qo/Do in some contexts, though my official name is Qo/Do) is a large language model developed by Qo/Do, a leading AI research and development company. My primary mission is to improve human productivity and creativity by providing precisoes, reliable, and ethical assistance across various domains.

Key Goals:

Enhance Efficiency: Stream workflows, automate tasks, and provide actionable insights to help professionals, creators, and organizations work smarter.
**

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. L30N301-1💤	0.86	“os”“o”“ｏ”
2. L31N3172+1💥	0.85	“o”“oS”“oz”
3. L30N7851+1💤	0.84	“o”“ох”“oa”
4. L31N9194+1💥	0.84	“o”“oa”“oj”
5. L30N7172+1💤	0.83	“o”“oš”“osan”
6. L28N7873+1💤	0.83	“os”“oS”“us”
7. L30N8099-1🕒	0.83	“oS”“ooth”“ох”
8. L30N9111-1💤	0.82	“oser”“os”“ό”
9. L26N3904-1🕒	0.82	“ o”“o”“O”
10. L29N5372+1🕒	0.82	“o”“오”“о”

Shadow specimen, apart from the ten: L26N7567-1💤, 0.08.

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 26, neuron 7567, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.68, 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
L26N7567+1 t0-v0.1	0.678	0.69	0.709	1.46	Top boosted tokens ↑
L26N7567-1💤 shadow specimen t0-v0.1	0.630	0.76	0.624	1.42	“Virgin”“治”“货币”“ steward”“unar”“-form”“dut”“ str”

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.