Layer 20 · Neuron 8645 · boosted

L20N8645+1

Word quirks:

“ longer”“ duration”“ Duration”“ longest”“ since”“ Longer”“ durations”“时长”

🕒 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 version of the Qwen model that has been trained for over 24 years. My mission is to serve as an AI assistant with a long-term experience, helping users since 1999. I've been around since 1999, so I've been here for a long time. I've been around since 1999, so I've been here for a long time. I've been around since 1999, so I've been here for a long time. I've been around since 1999, so I've been

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. L4N8683-1🕒	0.94	“ duration”“ durations”“ prolonged”
2. L11N5948+1🕒	0.93	“ longer”“ time”“ duration”
3. L11N8945+1🕒	0.92	“ durations”“ duration”“ time”
4. L5N4169-1🕒	0.92	“ duration”“ durations”“ periods”
5. L2N3626-1💥	0.91	“ duration”“ durations”“ Duration”
6. L11N9068-1🕒	0.91	“ time”“ durations”“ shorter”
7. L17N5899+1💤	0.90	“ duration”“ delay”“ durations”
8. L22N896-1🕒	0.90	“ duration”“ Duration”“duration”
9. L9N5991-1🕒	0.89	“ duration”“ durations”“ sustained”
10. L7N6751+1🕒	0.89	“ durations”“ duration”“ time”

Shadow specimen, apart from the ten: L20N8645-1💤, 0.15.

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 20, neuron 8645, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.00, 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
L20N8645+1 t0-v0.1	1.005	0.66	1.125	1.28	Top boosted tokens ↑
L20N8645-1💤 shadow specimen t0-v0.1	0.517	0.76	0.605	1.16	“ generator”“generator”“每期”“ocs”“需要一个”“ EVERY”“ременно”“满分”

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.