Layer 17 · Neuron 8220 · boosted

L17N8220+1

Word quirks:

“ continued”“ continue”“ continuing”“ continues”“继续”“continued”“ continuation”“ further”

💤 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 Tongyi Lab. My name is often shortened to "Qwen" or simply "Tongyi" (通义). I am an AI assistant designed to help with tasks ranging from creative writing and technical problem-solving to logical reasoning and code generation.

My mission is to provide accurate, helpful, and contextually relevant responses across diverse scenarios. Whether you need assistance with:

Technical queries (e.g., scientific facts, programming)
Creative tasks (e.g., stories, images, math problems)
**Practical applications

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. L10N8029-1🕒	1.00	“ continuation”“ continued”“ further”
2. L12N349-1🕒	1.00	“ continuing”“ continuation”“ continued”
3. L5N3849-1🕒	0.99	“ continuation”“ continue”“ continued”
4. L15N7819+1💤	0.99	“ continuation”“ continued”“ continuing”
5. L12N1489-1🕒	0.99	“ continued”“ continues”“ continuing”
6. L9N5940+1🕒	0.99	“ further”“ continued”“ continuing”
7. L11N4985+1🕒	0.99	“ continued”“ continuing”“ continuation”
8. L31N6372-1💥	0.99	“ continued”“continued”“ Continued”
9. L12N8127-1🕒	0.99	“ continuing”“ continued”“ continuation”
10. L18N6671-1💤	0.99	“ continued”“ continuing”“ continuation”

Shadow specimen, apart from the ten: L17N8220-1🕒, 0.40.

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 8220, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.57, top-1 agreement 0.73) 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
L17N8220+1 t0-v0.1	0.568	0.73	0.657	1.05	Top boosted tokens ↑
L17N8220-1🕒 shadow specimen t0-v0.1	1.144	0.60	1.416	1.41	“ longevity”“ lasts”“ linger”“ lifecycle”“ stays”“ once”“ lingering”“ lifetime”

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.