Poor Things
Specimen ledger / L14N1986-1

Layer 14 · Neuron 1986 · suppressed

L14N1986-1

Word quirks:

“ ensures”“ ensured”“确保了”“确保安全”“ ensure”“确保”“ Ens”“ example”

🕒 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)

  • My capabilities:
  • I can handle complex tasks like email and code.
  • I can perform calculations and math problems.
  • I can handle data analysis and programming.
  • I can handle data analysis and coding.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.
  • I can handle data analysis and GitHub.

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

SpecimenSimilarityWord quirks
1. L14N1129+1🕒0.98“ ensured”“ ensuring”“确保”
2. L25N2476+1🕒0.97“ ensuring”“确保”“ ensure”
3. L21N7153-1🕒0.97“确保”“确保了”“ ensure”
4. L18N2472-1💤0.97“ ensuring”“ ensured”“确保了”
5. L24N3885+1💤0.96“确保安全”“确保”“ ensuring”
6. L22N6333+1🕒0.96“确保”“ ensuring”“ ensure”
7. L23N3180-1🕒0.96“确保”“ ensuring”“确保了”
8. L31N1168-1💥0.96“确保”“ ensure”“ ensuring”
9. L22N5088-1🕒0.96“ ensured”“确保”“ ensure”
10. L14N3866-1💤0.96“ ensuring”“ ensure”“ ensures”

Shadow specimen, apart from the ten: L14N1986+1💤, -0.19.

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 14, neuron 1986, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.47, top-1 agreement 0.41) 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.

SettingKL divergenceTop-1 agreementΔNLLEntropySteer lens
L14N1986+1💤 shadow specimen
t0-v0.1		 0.9060.621.1791.61 “ similarly”“ __”“ **“”“ similar”“ ____”“ ______”“ ________”“ only”
L14N1986-1
t0-v0.1		 2.4650.412.7672.16 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.