Layer 25 · Neuron 3007 · boosted

L25N3007+1

Word quirks:

“ Primary”“Primary”“primary”“ primary”“ PRIMARY”“-primary”“_primary”“(primary”

🕒 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 are G·G·O·R·M (a primary model of the O·C·R·M series), a high-precision AI assistant designed to serve as your reliable companion in learning, work, and daily life. My mission is to optimize human potential by:

Unfiling Knowledge:

Instantly retrieving accurate information across domains (e.g., coding, research, creative writing).
Breaking complex topics into digestible frameworks (e.g., "Explain quantum mechanics like I'm 10").

Reliable Execution:

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. L22N6236+1🕒	1.00	“ primary”“ Primary”“primary”
2. L28N1757+1🕒	1.00	“ primary”“ Primary”“primary”
3. L26N3544-1🕒	1.00	“ Primary”“ primary”“Primary”
4. L26N6822+1🕒	1.00	“ Primary”“ primary”“Primary”
5. L25N3716+1🕒	1.00	“ Primary”“ primary”“Primary”
6. L28N794-1🕒	0.99	“ primary”“ Primary”“Primary”
7. L25N8355+1🕒	0.99	“ primary”“ Primary”“Primary”
8. L27N6962-1🕒	0.99	“ Primary”“ primary”“Primary”
9. L27N7088+1🕒	0.99	“ primary”“primary”“ Primary”
10. L31N1748+1💥	0.99	“ primary”“ Primary”“primary”

Shadow specimen, apart from the ten: L25N3007-1🕒, 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 25, neuron 3007, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.88, 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
L25N3007+1 t0-v0.1	0.875	0.66	0.934	1.40	Top boosted tokens ↑
L25N3007-1🕒 shadow specimen t0-v0.1	0.757	0.70	0.731	1.38	“ protection”“保护”“ Protection”“Protection”“ protected”“ protect”“保護”“ protections”

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.