Specimen ledger / L9N1913-1

Layer 9 · Neuron 1913 · suppressed

L9N1913-1

Word quirks:

“ corresponding”“ representing”“ analogous”“ similar”“ basically”“ analogy”“ represented”“ management”

🕒 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)

Here's a thinking process that will translate the concept of "mission" into a structured plan for a specific scenario:

Role: You are a creative strategist. Your name is Chen, and your mission is to guide a team through a complex crisis in a fictional setting.

Scenario: The situation involves a high-stakes emergency response involving a group of people (a team) who must navigate a dangerous environment to ensure safety. This mirrors the real-world application of crisis management principles in the context of the movie Interstellar (2014).

The Mission: In this scenario, the team consists of

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. Hairline Cracks✅ L12N8538-1	0.88	“ metaphor”“ analogy”“ analogous”
2. L14N6857-1💤	0.88	“ similar”“ similarly”“ corresponding”
3. L10N1274-1💥	0.88	“ analogous”“ similar”“ counterparts”
4. L12N4002+1🕒	0.86	“ corresponding”“ similar”“ similarly”
5. L14N7090+1🕒	0.85	“ analogy”“ analog”“ similarly”
6. L11N6124+1🕒	0.85	“ analogous”“ analogy”“ analog”
7. L10N4160-1🕒	0.85	“ similar”“ basically”“ referring”
8. L13N5369-1🕒	0.84	“ analogy”“ analogous”“ analog”
9. L14N6069-1💤	0.84	“ similarly”“ corresponding”“ analogous”
10. L14N7901+1🕒	0.84	“ representing”“ similar”“ analogous”

Shadow specimen, apart from the ten: L9N1913+1💤, -0.14.

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 9, neuron 1913, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.66, top-1 agreement 0.48) 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
L9N1913+1💤 shadow specimen t0-v0.1	0.740	0.69	0.897	1.22	“ “”“ ‘”“ “”“ ””“ ‘”“!”.”“ directly”“!””
L9N1913-1 t0-v0.1	1.656	0.48	1.966	1.77	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.