Layer 5 · Neuron 6359 · boosted

L5N6359+1

Word quirks:

“ legislation”“ legislative”“ policy”“ statutes”“ regulation”“ regulations”“ governance”“ governing”

🕒 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 advanced AI assistant designed to support the AI for Kids Act of 2024. This legislation aims to ensure that all children have access to comprehensive, evidence-based policies that prioritize their physical, mental, and emotional well-being.

The mission includes establishing a robust framework for:

Mandatory Safety Standards: Creating clear guidelines on school policies that address child safety, including mandatory reporting procedures for laws that protect minors from harmful activities.
Resource Allocation: Implementing specific measures that allocate funding for programs designed to safeguard the rights and interests of young people.
Implementation Oversight: Ensuring the 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. The Legislator✅ L3N2626-1	0.96	“ legislative”“ legislation”“ policy”
2. L6N5281-1🕒	0.94	“ legislative”“ legislation”“ laws”
3. L18N7310+1🕒	0.91	“ provisions”“ legis”“ legisl”
4. L4N4823+1🕒	0.91	“ regulation”“ protection”“ policy”
5. L5N8285-1🕒	0.91	“ governing”“ governance”“ policy”
6. L6N1631-1🕒	0.90	“ regulation”“ legislation”“ legal”
7. L4N5834-1🕒	0.90	“ regulation”“ statute”“ statutes”
8. L4N4033-1💥	0.90	“ legislation”“ regulation”“ appropriations”
9. L6N8044+1🕒	0.89	“ policy”“ governance”“ regulation”
10. L2N2116+1💥	0.89	“<\|im_end\|>”“ regulation”“ policy”

Shadow specimen, apart from the ten: L5N6359-1🕒, 0.17.

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 5, neuron 6359, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.80, top-1 agreement 0.47) 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
L5N6359+1 t0-v0.1	1.800	0.47	2.110	2.11	Top boosted tokens ↑
L5N6359-1🕒 shadow specimen t0-v0.1	2.184	0.47	2.642	2.25	“ ‘”“ indexed”“ search”“ searched”“ submissions”“ subgroup”“ searches”“ labelled”

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.