Layer 17 · Neuron 8154 · suppressed

L17N8154-1

Word quirks:

“ now”“ time”“ changed”“ previously”“ changing”“ later”“ commonly”“ formerly”

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

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. My primary mission is to be your AI assistant that helps you interact with information and solve problems across various domains. Here's how I aim to support you:

Answering questions: Providing accurate, concise, and context-aware responses for general knowledge, technical topics, or creative tasks.
Logical reasoning & problem-solving: Assisting in math, science, coding, and logical challenges (e.g., debugging code, analyzing data).
Multilingual & multimodal capabilities: Supporting 100+ languages

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. L13N5812+1💤	0.90	“ now”“ later”“ earlier”
2. L9N4995-1💤	0.87	“ previously”“ now”“ recently”
3. L13N3365-1💤	0.87	“ now”“ previously”“ recently”
4. L13N5318+1🕒	0.86	“ previously”“ now”“ earlier”
5. L14N6917+1🕒	0.85	“ previously”“ changing”“ constantly”
6. L11N7801-1💤	0.85	“ previously”“ now”“ formerly”
7. L15N6280-1💤	0.84	“ now”“ transition”“ changing”
8. L13N573+1💤	0.84	“ now”“ previously”“ formerly”
9. L11N6126-1🕒	0.83	“ now”“ balanced”“ nowadays”
10. L11N5350+1🕒	0.83	“ now”“ previously”“ still”

Shadow specimen, apart from the ten: L17N8154+1🕒, 0.25.

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 17, neuron 8154, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.55, top-1 agreement 0.71) 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
L17N8154+1🕒 shadow specimen t0-v0.1	1.006	0.61	1.303	1.52	“ initial”“ when”“ initialization”“ Initial”“ NEW”“ new”“ response”“/new”
L17N8154-1 t0-v0.1	0.548	0.71	0.676	1.33	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.