Specimen ledger / L5N2003-1

Layer 5 · Neuron 2003 · suppressed

L5N2003-1

Word quirks:

“ analysis”“ analyse”“ analysed”“ tracked”“ analyzing”“ analyzed”“ analys”“ analyses”

🕒 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 have carefully reviewed the transcript for S001E02 (Season 0 Episode 2). Here's a summary of the key points:

---

Current Status

Location: Space Station Lagrange (a strategic pause in the mission timeline).
Objective: The team is assessing progress on the current phase of the mission.
Key Players:
Commander: Leading the team with authority.
Engineer: Managing technical details and resource allocation.
Integrator: Coordinating efforts to ensure alignment with the

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. L10N4636+1🕒	0.99	“ analysis”“ analyses”“ analyzing”
2. L14N5511+1🕒	0.99	“ analysis”“ analyse”“ analyses”
3. L8N4784-1🕒	0.99	“ analysis”“ analys”“ analysed”
4. L10N6935+1💤	0.99	“ analysis”“ analyzed”“ analyses”
5. L8N1103+1🕒	0.98	“ analysis”“ analyzing”“ analyses”
6. L11N5923+1🕒	0.98	“ analysis”“ analyzed”“ analyzing”
7. L11N6767-1🕒	0.98	“ analysis”“ analyses”“ analys”
8. L9N8811+1🕒	0.98	“ analysis”“ analyzing”“ analyses”
9. L10N7523+1🕒	0.98	“ analysis”“ analyse”“ analysed”
10. L11N4239+1🕒	0.98	“ analysis”“ analyzed”“ analyzing”

Shadow specimen, apart from the ten: L5N2003+1💥, -0.11.

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 2003, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.77, top-1 agreement 0.45) 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
L5N2003+1💥 shadow specimen t0-v0.1	4.008	0.22	4.393	2.62	“<\|im_end\|>”“ _____”“ ____”“?\”“ ______”“____”“????”“_____”
L5N2003-1 t0-v0.1	1.766	0.45	2.104	2.24	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.