Layer 11 · Neuron 5623 · boosted

L11N5623+1

Word quirks:

“ answers”“ answer”“ answering”“ referenced”“ asking”“ answered”“ related”“ always”

🕒 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 an AI assistant. My mission is to provide helpful, accurate, and engaging responses to any query. Whether you're looking for information, assistance with a task, or simply want to chat, I'm here to support you!

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. L10N1842-1🕒	0.91	“ answers”“ answer”“ answered”
2. L9N2229-1💤	0.89	“ answer”“ answers”“ question”
3. L11N4238-1💤	0.87	“ answer”“ always”“ answers”
4. L11N2359-1🕒	0.87	“ answer”“ question”“ answers”
5. L7N4541+1💤	0.86	“ question”“ answer”“ questions”
6. L13N2177-1🕒	0.86	“ answer”“ answers”“ question”
7. L9N491+1🕒	0.86	“ response”“ answer”“ responses”
8. L16N1003+1🕒	0.85	“ answer”“ answers”“ answering”
9. L18N2206+1💤	0.85	“ answered”“ answering”“ answer”
10. L12N6094+1🕒	0.85	“ answer”“ answers”“ unanswered”

Shadow specimen, apart from the ten: L11N5623-1💤, 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 11, neuron 5623, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.85, top-1 agreement 0.63) 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
L11N5623+1 t0-v0.1	0.852	0.63	1.069	1.41	Top boosted tokens ↑
L11N5623-1💤 shadow specimen t0-v0.1	0.577	0.68	0.742	1.28	“ now”“ bigger”“ aligned”“ needed”“ ____”“ huge”“ focus”“ already”

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.