Layer 12 · Neuron 6874 · boosted

L12N6874+1

Word quirks:

“ specific”“ particular”“ **”“ **"”“ specifically”“ ***”“ *”“ __”

🕒 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, and my mission is to provide helpful, accurate, and safe information. I am designed to assist users in tasks that require human-like intelligence and decision-making. My role is to offer support in areas where I can make a positive difference, such as answering questions about science, technology, or general knowledge. I aim to be a reliable source of information that users can trust.

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. L15N3146+1🕒	0.89	“ specifically”“,**”“ specific”
2. L10N8266-1🕒	0.88	“ specially”“ specific”“ ***”
3. L11N6741-1🕒	0.87	“ specific”“ ”“ "”
4. L11N8216+1🕒	0.86	“ **"”“ specific”“ specifically”
5. L11N4876+1💤	0.86	“ *”“ ”“ ***”
6. L10N1107-1💤	0.85	“ **”“ specific”“ exact”
7. L8N3727-1💤	0.85	“ specific”“ ”“ *”
8. L11N9017-1🕒	0.84	“ ”“ “”“ **"”
9. L8N5498+1💤	0.84	“ specifically”“ ”“ *”
10. L10N9153+1🕒	0.84	“ specific”“ "”“ ”

Shadow specimen, apart from the ten: L12N6874-1💤, -0.07.

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 6874, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.96, top-1 agreement 0.60) 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
L12N6874+1 t0-v0.1	0.955	0.60	1.237	1.32	Top boosted tokens ↑
L12N6874-1💤 shadow specimen t0-v0.1	0.757	0.62	0.910	1.03	“ ⏎”“ Prepare”“ prepare”“ indication”“ preparation”“ analyse”“ prep”“ prepared”

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.