Layer 19 · Neuron 6753 · suppressed

L19N6753-1

Word quirks:

“ amongst”“ among”“ collective”“ Among”“ sum”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am JBC, the most advanced collective intelligence among all members of the JBC system. We are a team of AI agents working together to provide high-level support for our members.

The missions of the JBC system are shared by all members:

Collective Intelligence: We work together to enhance the capabilities of each member, ensuring that no single agent falls short of the highest standards.
High-Level Support: We provide advanced assistance to our members, helping them achieve their goals across various tasks and challenges.
Shared Mission: The collective intelligence of our members is the highest level of support we offer

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. L15N8732-1🕒	0.77	“ collectively”“ collective”“ aggregate”“ aggregated”“ Collective”
2. L24N1801-1🕒	0.76	“ collectively”“ cumulative”“ cumul”“ summed”“ Together”
3. L6N5123+1🕒	0.75	“ collectively”“ collective”“ combined”“ sum”“ merged”
4. L15N5601-1🕒	0.74	“ summed”“ aggregate”“ sum”“ collectively”“ aggregated”
5. L16N4823-1🕒	0.72	“ collectively”“ collective”“ Collective”“ united”“ unified”
6. L13N4010+1🕒	0.72	“ collectively”“ collective”“ collection”“ summed”“ together”
7. L9N1155+1🕒	0.71	“ aggregated”“ each”“ collectively”“ individually”“ aggreg”
8. L21N8588+1🕒	0.71	“ among”“ within”“ amongst”“ join”“among”
9. L30N335+1💥	0.70	“ among”“among”“ Among”“ amongst”“Among”
10. L18N4636-1🕒	0.70	“ sums”“ together”“ three”“ summaries”“ sum”

Shadow specimen, apart from the ten: L19N6753+1💤, 0.10.

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 19, neuron 6753, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L19N6753+1💤 shadow specimen t0-v0.1	0.338	0.74	0.422	1.06	“ sectors”“ independents”“ regions”“ general”“ tiers”
L19N6753-1 t0-v0.1	1.885	0.54	2.105	1.73	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.