Paper Citation Record · LEDGER

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance

As of 19 August 2026, this Paper Citation Record lists 37 of 37 outbound references and 0 inbound Pith citation observations for arXiv:2607.19386.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2607.19386 v1

Coverage vector

measured 37 of 37 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-15T15:40:27.706679Z

measured 37 of 37 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-19T06:32:44.657259+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: cited_works

Reference resolution

37 of 37 outbound references displayed

verified exact0
verified fuzzy23
unresolved13
parse uncertain0
malformed identifier1
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation fc2acce6-e64e-404b-b50b-54bfd4ce18ca · outbound

This paper cites Pythia: a suite for analyzing large language models across training and scaling.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Pythia: a suite for analyzing large language models across training and scaling

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.562353Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d5ac0dbc-c8ba-446f-85e1-6a4309621833 · outbound

This paper cites Language models can explain neurons in language models, 2023.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Language models can explain neurons in language models, 2023

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.567155Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 62ebb96b-dd1c-4fb2-a426-ecc75b34dbb3 · outbound

This paper cites Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread, 2023.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread, 2023

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.571356Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5d797047-80a2-4976-806a-58daa5f791d8 · outbound

This paper cites Learning multi-level features with matryoshka sparse autoencoders.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Learning multi-level features with matryoshka sparse autoencoders

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.343227Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation f56b1e10-1f41-4f71-83e1-b1bb8170c523 · outbound

This paper cites Improving Steering Vectors by Targeting Sparse Autoencoder Features.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Improving Steering Vectors by Targeting Sparse Autoencoder Features

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.580176Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae5621bf-2ff7-4fb1-81f8-e70b7b795ef8 · outbound

This paper cites Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.584739Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e07126c9-a1dc-41df-b022-981aeb867eac · outbound

This paper cites A is for absorption: Studying feature splitting and absorption in sparse autoencoders, 2024.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance A is for absorption: Studying feature splitting and absorption in sparse autoencoders, 2024

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.329969Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 3b1ef15a-4b23-40e1-b7dd-a3812998721f · outbound

This paper cites GitHub Code dataset.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance GitHub Code dataset

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.315903Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation e35d06a0-8b33-4cba-abf7-1a03c349565f · outbound

This paper cites Sparse autoencoders find highly interpretable features in language models.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sparse autoencoders find highly interpretable features in language models

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.302703Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 17051944-f72f-431d-bc3f-8b51532c7a39 · outbound

This paper cites Wikimedia downloads.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Wikimedia downloads

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.601068Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6297f170-8a63-4fa6-85b6-ff2d6ad123a5 · outbound

This paper cites The pile: An 800gb dataset of diverse text for language modeling, 2020.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance The pile: An 800gb dataset of diverse text for language modeling, 2020

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.280432Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 0fbfe632-ccda-41e1-bd6e-251bb21448e2 · outbound

This paper cites Scaling and evaluating sparse autoencoders.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Scaling and evaluating sparse autoencoders

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.608736Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d84049c3-301d-4c0e-a920-da98c0da9c52 · outbound

This paper cites Apertus: Democratizing open and compliant LLMs for global language environments, 2025.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Apertus: Democratizing open and compliant LLMs for global language environments, 2025

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.257161Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation df779e8f-9b3b-458c-978a-923652d03c3f · outbound

This paper cites Open source automated interpretability for sparse autoencoder features.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Open source automated interpretability for sparse autoencoder features

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.243501Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 69f21c61-a0b1-43c3-a85e-c0a35d17b2a2 · outbound

This paper cites Are sparse autoencoders useful? a case study in sparse probing.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Are sparse autoencoders useful? a case study in sparse probing

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.230328Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation f43aca0f-b6b4-4393-be30-829f3b120907 · outbound

This paper cites SAEBench: A comprehensive benchmark for sparse autoencoders in language model interpretability.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance SAEBench: A comprehensive benchmark for sparse autoencoders in language model interpretability

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.216544Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 2f3d4cca-dab2-40ec-9d04-ae93178bf7f2 · outbound

This paper cites A guideline of selecting and reporting intraclass correlation coefficients for reliability research.Journal of chiropractic medicine, 15(2):155–163, 2016.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance A guideline of selecting and reporting intraclass correlation coefficients for reliability research.Journal of chiropractic medicine, 15(2):155–163, 2016

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.201654Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 07f91b51-f27d-4f21-bb00-7538f3bce9c7 · outbound

This paper cites Sanity checks for sparse autoencoders: Do saes beat random baselines?arXiv preprint arXiv:2602.14111, 2026.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sanity checks for sparse autoencoders: Do saes beat random baselines?arXiv preprint arXiv:2602.14111, 2026

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.631351Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 941be90c-aee5-4684-ac00-8d9f429c9e64 · outbound

This paper cites A stability index for feature selection.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance A stability index for feature selection

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.188387Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation fbdc88b6-edff-4d4a-a9ae-5c4c3aa67218 · outbound

This paper cites Li, Suraj Srinivas, Usha Bhalla, and Himabindu Lakkaraju.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Li, Suraj Srinivas, Usha Bhalla, and Himabindu Lakkaraju

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.174309Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation e503bf00-f334-4dcb-818a-7cb8685d4842 · outbound

This paper cites Intraclass correlation–a discussion and demonstration of basic features.PloS one, 14(7):e0219854, 2019.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Intraclass correlation–a discussion and demonstration of basic features.PloS one, 14(7):e0219854, 2019

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.160935Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation bcd7eeb3-eda2-4972-8f47-f6f34e51b835 · outbound

This paper cites Towards principled evaluations of sparse autoencoders for interpretability and control.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Towards principled evaluations of sparse autoencoders for interpretability and control

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.147104Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 9abdb850-490a-41fd-bdd4-7c303b55324d · outbound

This paper cites Sparse feature circuits: Discovering and editing interpretable causal graphs in language models.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sparse feature circuits: Discovering and editing interpretable causal graphs in language models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.650368Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7936250f-5973-4139-9eb9-f950b8c4b1ec · outbound

This paper cites Rethinking evaluation of sparse autoencoders through the representation of polysemous words.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Rethinking evaluation of sparse autoencoders through the representation of polysemous words

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.122513Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation ae537b66-192e-4713-aea7-beacc0a34a3e · outbound

This paper cites an unresolved cited work.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Unresolved cited work

Reference 25

Resolution: unresolved; raw_fallback, observed 2026-08-15T15:40:28.108417Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 4829db98-393d-4bd4-ab38-b952ceef7d61 · outbound

This paper cites Sparse autoencoders trained on the same data learn different features.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sparse autoencoders trained on the same data learn different features

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.094034Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 9862a3fe-97a4-4dd2-a0cf-f378896bab05 · outbound

This paper cites Automatically interpreting millions of features in large language models.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Automatically interpreting millions of features in large language models

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.080121Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 7acceead-2e40-473e-b227-6e4aa46a417f · outbound

This paper cites FADE: Why bad descriptions happen to good features.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance FADE: Why bad descriptions happen to good features

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.065957Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 773388a2-99ce-4343-8bc0-458daedca425 · outbound

This paper cites an unresolved cited work.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Unresolved cited work

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.673178Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation babfafae-fa97-441f-885f-883a5bf48692 · outbound

This paper cites Improving sparse decomposition of language model activations with gated sparse autoencoders.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Improving sparse decomposition of language model activations with gated sparse autoencoders

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.042878Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation c81f4395-718f-4dea-95ed-a0c3aabb5590 · outbound

This paper cites Sullivan and Richard Feinn.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Sullivan and Richard Feinn

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.029064Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation e490a44f-dc28-45d7-aafc-7293adcdade6 · outbound

This paper cites an unresolved cited work.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Unresolved cited work

Reference 32

Resolution: unresolved; raw_fallback, observed 2026-08-15T15:40:28.014963Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation adcb075c-e4fd-4787-ab96-6727b882271c · outbound

This paper cites Daniel Freeman, Theodore R.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Daniel Freeman, Theodore R

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:28.000854Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation ae010ecc-adc9-482c-9391-9ba0a4ddaed4 · outbound

This paper cites Persona features control emergent misalignment.arXiv preprint arXiv:2506.19823, 2025.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Persona features control emergent misalignment.arXiv preprint arXiv:2506.19823, 2025

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-15T15:40:27.693496Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f6b05b1f-4c5c-4752-9b62-a6b667d40285 · outbound

This paper cites Axbench: Steering LLMs? even simple base- lines outperform sparse autoencoders.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Axbench: Steering LLMs? even simple base- lines outperform sparse autoencoders

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:27.987829Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation bb8e841d-c233-42a9-94ba-17ae32439ac2 · outbound

This paper cites Mathematical equations and formulas.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Mathematical equations and formulas

Reference 36

Resolution: malformed identifier; raw_fallback, observed 2026-08-15T15:40:27.974427Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Observation 3ad60b9a-5913-4bcd-be77-d0b748173032 · outbound

This paper cites Justification: There is no crowdsourcing nor research with human subjects.

Building Fast, Evaluating Slow: Pipeline Choices Dominate Autointerpretability Score Variance Justification: There is no crowdsourcing nor research with human subjects

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T15:40:27.960201Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-19T06:32:44.657259+00:00.

Pith citing papers

No inbound Pith citation observations are available.