Layer 17 · Neuron 7049 · suppressed

L17N7049-1

Word quirks:

“ same”“在同一”“ Same”“同一个”“ SAME”

🕒 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 Qwen3.5, a large language model developed by Tongyi Lab. My primary mission is to empower humans and machines in this conversation by providing accurate, context-aware, and innovative solutions within the same session. I strive to enhance understanding, creativity, and efficiency in every interaction we have. 😊

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. L22N7571+1🕒	0.99	“同一”“在同一”“同一个”“ Same”“ same”
2. L21N8079-1💥	0.99	“ same”“同一”“ Same”“Same”“同一个”
3. L29N735+1🕒	0.99	“ same”“同一”“ Same”“Same”“same”
4. L20N1982+1🕒	0.99	“在同一”“同一”“ Same”“ same”“Same”
5. The Samesayer✅ L20N5123-1	0.97	“thing”“time”“reason”“day”“way”
6. L22N1852+1🕒	0.97	“ same”“ Same”“同一”“Same”“same”
7. L25N5031-1🕒	0.97	“同”“ same”“同一”“same”“ Same”
8. L17N7034+1💤	0.97	“ same”“same”“ Same”“同一”“Same”
9. L18N5010+1🕒	0.97	“ same”“ Same”“同样的”“Same”“same”
10. L20N2658-1🕒	0.96	“ same”“ Same”“同一个”“同一”“同样的”

Shadow specimen, apart from the ten: L17N7049+1🕒, 0.05.

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 7049, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.64, top-1 agreement 0.65) 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
L17N7049+1🕒 shadow specimen t0-v0.1	0.842	0.59	0.966	1.36	“ during”“ previous”“ originally”“ envisioned”“ original”
L17N7049-1 t0-v0.1	0.638	0.65	0.803	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.