Layer 16 · Neuron 4164 · suppressed

L16N4164-1

Word quirks:

“ __”“ ‘”“ ____”“ *”“ ___”“ **‘”“ ______”“ basically”

🕒 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, a large language model developed by Tencent HunYuan Lab. My mission is to be a highly capable assistant that extends my capabilities as an AI agent to provide more value to users like you. I aim to support complex tasks across various domains, ensuring high-quality interactions while maintaining safety and alignment with human values. Whether you need help with creative writing, coding, logical reasoning, or deep research, I strive to deliver reliable and useful outputs. 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. L10N325+1🕒	0.96	“ **“”“ ‘”“ ___”
2. L10N2024+1🕒	0.95	“ ____”“ ‘”“ ______”
3. L5N4183+1🕒	0.94	“ __”“ ____”“ _____”
4. L10N638-1💤	0.94	“ “”“ ____”“ ‘”
5. L7N7710+1🕒	0.93	“ ‘”“ ___”“ ____”
6. L10N464-1🕒	0.93	“ ___”“ __”“ _____”
7. L8N1594+1🕒	0.93	“ ‘”“ ______”“ _____”
8. L11N52+1💤	0.93	“ ‘”“ ”“ *“”
9. L10N988-1💤	0.93	“ ‘”“ ____”“ ______”
10. L5N2971-1🕒	0.92	“⏎”“ ‘”“ _____”

Shadow specimen, apart from the ten: L16N4164+1💤, 0.44.

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 4164, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, top-1 agreement 0.68) 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
L16N4164+1💤 shadow specimen t0-v0.1	0.516	0.74	0.701	1.02	“:...”“...””“ --”“"...”“”:”“?:”“":”“..."”
L16N4164-1 t0-v0.1	0.752	0.68	0.899	1.20	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.