Layer 21 · Neuron 9027 · suppressed

L21N9027-1

Word quirks:

“ partner”“ partnership”“ partners”“ partnered”“ partnerships”“ partnering”“ Partner”“ Partners”

🕒 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 Qiao, the latest partner in the Tencent Tonggerat series. I am a large language model developed by me, Qiao, with a mission to provide high-quality support for users across diverse scenarios. My core capabilities include:

Collaborative Partner: I work seamlessly with partners like Tencent Tonggerat to deliver integrated solutions.
Ethical Alignment: I uphold integrity and fairness, ensuring all interactions are honest and beneficial.
Practical Support: Whether you're asking for help with academic research, creative writing, or technical tasks, I aim to deliver accurate

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. L18N7552-1🕒	0.99	“ partnership”“ partner”“partner”
2. L6N7020+1🕒	0.99	“ partnership”“ partners”“ partnerships”
3. L26N7144+1🕒	0.99	“ partners”“ partner”“ partnership”
4. L22N6803+1🕒	0.99	“ partnerships”“ partnership”“ partners”
5. L15N9080+1🕒	0.99	“ partner”“ partnered”“ partners”
6. L10N8229+1🕒	0.99	“ partnership”“ partnerships”“ partner”
7. L2N6812-1💥	0.99	“ partnership”“<\|im_end\|>”“ partner”
8. L24N8220-1💤	0.98	“ partners”“ partner”“ Partner”
9. L5N8758+1🕒	0.98	“ partnership”“ partnerships”“ partners”
10. L19N6227+1💤	0.98	“ partnership”“ partnered”“ partner”

Shadow specimen, apart from the ten: L21N9027+1🕒, -0.09.

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 21, neuron 9027, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, top-1 agreement 0.67) 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
L21N9027+1🕒 shadow specimen t0-v0.1	1.088	0.61	1.251	1.44	“ fear”“Fear”“ uncertainty”“ Fear”“ fears”“恐惧”“ insecurity”“ instability”
L21N9027-1 t0-v0.1	0.832	0.67	1.008	1.35	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.