Layer 29 · Neuron 5524 · suppressed

L29N5524-1

Word quirks:

“市场”“ Market”“Market”“markets”“市场的”“ Markets”“-market”“market”

🕒 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 Qwen3.5, the latest large-scale foundation model developed by Alibaba Cloud's Market-Driven AI (Market-Driven AI) team. My primary mission is to empower users with advanced, reliable, and accessible AI capabilities across diverse scenarios. Here's how I aim to deliver on this mission:

Core Objectives:

Seamless Integration:

Support natural, fluid interactions through ultra-advanced natural language understanding and multi-turn context management, ensuring accurate responses even in complex or ambiguous queries.

Market-Driven Innovation:

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. L23N8037-1🕒	1.00	“ markets”“ market”“ Markets”
2. L26N1653+1🕒	0.99	“ markets”“ market”“ Markets”
3. One-Track Market✅ L14N5751-1,L25N2343-1	0.99	“市场”“ market”“ markets”
4. L24N5495-1🕒	0.99	“ market”“ Market”“Market”
5. L31N4044+1💥	0.99	“ market”“ Market”“市场”
6. L26N7377+1🕒	0.98	“ market”“市场”“ Market”
7. L29N6038-1🕒	0.98	“ market”“市场”“market”
8. L23N7121+1🕒	0.98	“ market”“ Market”“market”
9. L25N4215-1🕒	0.98	“market”“市场”“ market”
10. L20N6117+1🕒	0.98	“ market”“ Market”“ markets”

Shadow specimen, apart from the ten: L29N5524+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 29, neuron 5524, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.35, top-1 agreement 0.79) 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
L29N5524+1🕒 shadow specimen t0-v0.1	0.330	0.79	0.400	1.64	“usi”“anean”“gar”“宪”“пі”“plane”“PIO”“叶子”
L29N5524-1 t0-v0.1	0.351	0.79	0.435	1.54	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.