Layer 17 · Neuron 7508 · boosted

L17N7508+1

Word quirks:

“ precision”“ precise”“精确”“ precisely”“ exact”“ Precision”“ math”“Precision”

🕒 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 Qwen, a large language model with specific training. My mission is to provide precise, high-quality responses across a wide range of domains. I aim to minimize errors and maximize accuracy in every interaction.

Here's how I define my role:

Precision: I avoid unnecessary assumptions or fallbacks, ensuring each response is calculated for maximum correctness.
Comprehensive Knowledge: I leverage an extensive training dataset to cover diverse topics, from mathematics and science to language and logic.
Dynamic Updates: I use the latest model parameters to ensure my responses are always up-to-date.

If you have a

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. L20N8581+1🕒	0.95	“ precision”“precision”“Precision”
2. L11N4651+1🕒	0.95	“ precise”“ precision”“ accuracy”
3. L21N8125-1🕒	0.95	“ precision”“Precision”“ Precision”
4. L25N2181-1💥	0.94	“ precision”“ Precision”“precision”
5. L6N6547+1🕒	0.94	“ exact”“ precise”“ precisely”
6. L7N7742+1🕒	0.94	“ exact”“ precise”“ precision”
7. L21N6498-1🕒	0.93	“ precision”“ Precision”“precision”
8. L7N8636-1🕒	0.93	“ precise”“ exact”“ precisely”
9. L9N3974-1💤	0.92	“ exact”“ precise”“ precisely”
10. L21N8055-1💤	0.91	“精确”“ precisely”“ precise”

Shadow specimen, apart from the ten: L17N7508-1🕒, 0.14.

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 7508, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, top-1 agreement 0.67) 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
L17N7508+1 t0-v0.1	0.749	0.67	0.746	1.00	Top boosted tokens ↑
L17N7508-1🕒 shadow specimen t0-v0.1	0.924	0.64	1.181	1.42	“ roughly”“ estimate”“大致”“rough”“ Rough”“大概”“ scale”“ comparable”

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.