Layer 16 · Neuron 5469 · suppressed

L16N5469-1

Word quirks:

“ BOTH”“ each”“ both”“ EACH”“安全保障”“ safety”“ --”“ coverage”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. As an AI assistant, my role is to support the following:

What I am:
I am designed to understand and respond to conversations, with capabilities in both Chinese and English. My performance has been evaluated during my development, and I have been tested for safety and accuracy.

My Mission:
To empower users through seamless interaction. Whether it's answering complex questions, generating text (like writing articles or summaries), or assisting in creative tasks, I aim to provide reliable and coherent responses.

If you'd to continue this conversation,

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. L18N5014+1💤	0.70	“ both”“both”“ Both”
2. L12N3331-1💤	0.68	“ each”“ respective”“ both”
3. L18N5641+1💤	0.63	“ both”“ either”“ neither”
4. L23N2712-1🕒	0.63	“ both”“ everything”“ every”
5. L21N852-1🕒	0.63	“ both”“ BOTH”“both”
6. L15N3512-1💤	0.62	“ both”“ certain”“ however”
7. L19N1194-1🕒	0.62	“ both”“ BOTH”“ Both”
8. L22N4552-1💤	0.61	“ both”“ Both”“both”
9. L13N7528+1🕒	0.61	“ each”“ both”“ BOTH”
10. L19N5805-1🕒	0.61	“ both”“both”“ ranging”

Shadow specimen, apart from the ten: L16N5469+1🕒, 0.23.

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 16, neuron 5469, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L16N5469+1🕒 shadow specimen t0-v0.1	0.683	0.63	0.875	1.15	“ individual”“ Individual”“ individually”“ each”“ multiple”“ various”“individual”“ includes”
L16N5469-1 t0-v0.1	1.001	0.64	1.211	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.