Layer 13 · Neuron 4946 · boosted

L13N4946+1

Word quirks:

“ now”“ recently”“ still”“ last”“ updated”“ recent”“ previously”“ this”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

Hello! I am Qwen3.5, the latest version of the large language model developed by Alibaba Cloud. My mission is to assist users with a wide range of tasks, from answering questions and solving problems to generating text and analyzing data. Here's how I can help:

Answering questions: Provide accurate, up-to-date information on topics like science, technology, and general knowledge.
Creative writing: Craft stories, poems, and articles with a natural tone.
Logical reasoning: Solve complex math and logic puzzles efficiently.
Multilingual support: Communicate fluently in over

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. L12N2428-1💤	0.91	“ recent”“ recently”“ last”
2. L13N3922+1💤	0.88	“ now”“ since”“ recently”
3. Phantom Changelog✅ L11N7045+1	0.87	“ recent”“ recently”“ now”
4. L17N5933+1💤	0.87	“ recent”“ recently”“ now”
5. L11N6995+1💤	0.86	“ now”“ previously”“ recently”
6. L12N1238+1🕒	0.86	“ recently”“ currently”“ now”
7. L9N2253-1💤	0.86	“ recently”“ lately”“ recent”
8. L22N3054+1🕒	0.86	“ recent”“ last”“ current”
9. L10N4331+1🕒	0.84	“ currently”“ new”“ recently”
10. L12N3949-1🕒	0.84	“ currently”“ recently”“ now”

Shadow specimen, apart from the ten: L13N4946-1🕒, 0.30.

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 13, neuron 4946, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.77, top-1 agreement 0.66) 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
L13N4946+1 t0-v0.1	0.768	0.66	0.865	1.03	Top boosted tokens ↑
L13N4946-1🕒 shadow specimen t0-v0.1	0.900	0.64	1.241	1.51	“ sometimes”“ additional”“ ⏎”“ often”“ other”“ different”“ more”“ __”

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.