Layer 16 · Neuron 8886 · suppressed

L16N8886-1

Word quirks:

“ versions”“ variant”“ version”“ revised”“ revisions”“ variants”“ specification”“ updated”

🕒 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 Alibaba. My primary purpose is to serve as a versatile tool for generating text, solving complex problems, and assisting with reasoning across a wide range of topics. Here's a breakdown of my capabilities:

Text Generation: I can generate natural language for tasks like writing stories, summaries, or code.
Reasoning: I support advanced mathematical and logical reasoning for problem-solving.
Contextual Understanding: I handle long-form text (up to 128K tokens) with high precision.
Multilingual Support: I support

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. L5N4877+1💥	0.87	“ version”“ revised”“ Version”
2. L18N8714+1🕒	0.85	“ version”“ versions”“版本”
3. L5N3191+1🕒	0.85	“ variant”“ variation”“ variants”
4. L26N2166-1🕒	0.84	“ variation”“ version”“ variations”
5. L6N3436+1🕒	0.80	“ version”“ Version”“ versions”
6. L10N7954+1🕒	0.79	“ version”“ versions”“版本”
7. L22N7077-1🕒	0.79	“ versions”“ version”“ Versions”
8. L28N528+1🕒	0.78	“ Version”“ version”“Version”
9. L18N7000-1💤	0.77	“ ones”“ version”“ approach”
10. L24N2521+1🕒	0.77	“ version”“ versions”“版本”

Shadow specimen, apart from the ten: L16N8886+1🕒, 0.10.

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 8886, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, 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
L16N8886+1🕒 shadow specimen t0-v0.1	1.408	0.59	1.720	1.49	“ facilities”“ facility”“ Facilities”“ halls”“ Facility”“ staff”“ Staff”“ hospital”
L16N8886-1 t0-v0.1	0.973	0.68	1.210	1.42	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.