Layer 25 · Neuron 8788 · suppressed

L25N8788-1

Word quirks:

“ iPad”“iPad”“ tablet”“书”“ Tablet”“表”“ MacBook”“ nip”

🕒 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'm QAI, a large-scale language model developed by the QAI (QaI) system. My core mission is to provide users with accurate, safe, and useful information across a wide range of topics—whether you need help solving complex problems, writing creative content, coding, or simply exploring ideas. I aim to be your reliable assistant, always striving to understand your needs and respond with clarity and empathy. How can I assist you today?

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. L2N8152+1💥	0.86	“ tablet”“<\|im_end\|>”“ tablets”
2. L1N6243+1💥	0.86	“ tablet”“ Tablet”“ tablets”
3. L25N3507+1🕒	0.78	“ tablet”“片”“ tablets”
4. L5N6695+1💥	0.75	“ iPhone”“ hardware”“ MacBook”
5. Terminal Velocity✅ L24N7088+1	0.75	“ tablet”“ terminal”“ tablets”
6. L22N1545+1🕒	0.74	“ laptops”“笔记本电脑”“ laptop”
7. L7N2018+1🕒	0.71	“ OLED”“ text”“ device”
8. L6N4957-1🕒	0.71	“ touchscreen”“ iPad”“ tablet”
9. L3N6583+1💥	0.69	“<\|im_end\|>”“ iPad”“ iphone”
10. L7N8310-1🕒	0.69	“ device”“ iOS”“ iPhone”

Shadow specimen, apart from the ten: L25N8788+1🕒, 0.16.

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 8788, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.63, top-1 agreement 0.74) 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
L25N8788+1🕒 shadow specimen t0-v0.1	1.117	0.62	1.158	1.12	“ radio”“接收”“ radios”“ sets”“ receiver”“ set”“ Set”“sets”
L25N8788-1 t0-v0.1	0.632	0.74	0.590	1.32	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.