Layer 15 · Neuron 2497 · boosted

L15N2497+1

Word quirks:

“ previously”“ previous”“ implied”“ absent”“ implicit”“ implicitly”“ initially”“ absence”

💤 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 core mission is to provide practical assistance and support for users across various tasks, including complex reasoning, content creation, and code generation. I aim to enhance user experience through natural interactions and efficient problem-solving. If you have specific needs or questions, feel free to ask!

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. L14N5638-1🕒	0.77	“ implicit”“ implicitly”“ implied”
2. L13N5020-1💤	0.76	“ without”“<\|im_end\|>”“ earlier”
3. L8N2808-1🕒	0.75	“ no”“ indirectly”“ implied”
4. L17N8825-1🕒	0.74	“ implicit”“ implicitly”“ unless”
5. The Neverborn✅ L15N7859-1	0.74	“ implicitly”“ implicit”“ once”
6. L16N8472+1🕒	0.73	“ implicit”“ implicitly”“ unseen”
7. L15N2456+1💤	0.72	“ hidden”“ implicitly”“ unknown”
8. L13N7338+1💤	0.72	“ implicitly”“ implicit”“ indirectly”
9. L15N8897-1🕒	0.72	“ implicitly”“ implicit”“ yourself”
10. L17N7772+1🕒	0.71	“ implicit”“ assumed”“ implied”

Shadow specimen, apart from the ten: L15N2497-1🕒, -0.00.

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 15, neuron 2497, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.58, top-1 agreement 0.74) 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
L15N2497+1 t0-v0.1	0.578	0.74	0.754	1.28	Top boosted tokens ↑
L15N2497-1🕒 shadow specimen t0-v0.1	0.678	0.68	0.877	1.13	“ *”“ "”“ these”“ "”“ maybe”“ sometimes”“ “”“ ⏎⏎”

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.