Layer 12 · Neuron 5209 · suppressed

L12N5209-1

Word quirks:

“ only”“ ONLY”“ single”“ *”“ partial”“-only”“ unified”“ acknowledgment”

🕒 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 a large language model, and my mission is to provide helpful, accurate, and context-aware assistance. Whether you're asking for a quick fact check, complex problem-solving, or creative collaboration, I'm designed to adapt to your needs—just like a conversation that should feel natural. For instance, I can offer tailored responses based on your specific questions, while also ensuring that the information I provide aligns with your intent and expertise level.

---

Example of how I might respond:
"Could you explain this concept in simpler terms?" → A focused answer that assumes prior knowledge.
*"What about the reverse?"

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. L12N7997+1🕒	0.87	“ only”“ ONLY”“ Only”
2. L16N2959-1💤	0.87	“ only”“ ONLY”“ sole”
3. L17N7487+1🕒	0.86	“ only”“-only”“only”
4. L13N4799+1🕒	0.86	“ only”“-only”“ ONLY”
5. L16N1491-1💤	0.86	“ only”“ solely”“ ONLY”
6. L14N6165+1💤	0.86	“ only”“ ONLY”“only”
7. L17N4537+1💤	0.86	“ only”“ just”“only”
8. L10N1368+1🕒	0.85	“ only”“ ONLY”“ alone”
9. L16N9208+1💤	0.85	“ ONLY”“-only”“ sole”
10. L11N5057+1🕒	0.85	“ only”“ ONLY”“-only”

Shadow specimen, apart from the ten: L12N5209+1🕒, 0.19.

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 12, neuron 5209, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.87, top-1 agreement 0.64) 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
L12N5209+1🕒 shadow specimen t0-v0.1	1.166	0.61	1.453	1.25	“ sometimes”“ usually”“ often”“ always”“ each”“ typically”“ then”“ before”
L12N5209-1 t0-v0.1	0.866	0.64	1.052	1.23	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.