Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent | Paper Citation Record

Observation bf0ea858-5a82-4e7a-a4b5-1d78bd4d51bd · outbound

This paper cites Rossi, Varun Manju- natha, Roshan Santhosh, Ruiyi Zhang, Soheil Feizi, and Nedim Lipka.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Rossi, Varun Manju- natha, Roshan Santhosh, Ruiyi Zhang, Soheil Feizi, and Nedim Lipka

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.285944Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8a0931e4-8ffa-4800-8811-6c16431e45ac · outbound

This paper cites Probing classifiers: Promises, shortcomings, and advances.Com- putational Linguistics, 48(1):207–219, March 2022.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Probing classifiers: Promises, shortcomings, and advances.Com- putational Linguistics, 48(1):207–219, March 2022

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.340677Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 92b6e277-0e6d-4683-a501-0a35b38b4a7a · outbound

This paper cites Language models can ex- plain neurons in language models.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Language models can ex- plain neurons in language models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.482176Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 45cb9374-6c3d-4806-84d7-b1f7d3fe6457 · outbound

This paper cites Burke, Tristan Hume, Shan Carter, Tom Henighan, and Christopher Olah.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Burke, Tristan Hume, Shan Carter, Tom Henighan, and Christopher Olah

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.685248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d6c1786e-c4ef-4bab-8967-01781a552c16 · outbound

This paper cites Building and evaluating alignment auditing agents.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Building and evaluating alignment auditing agents

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.876425Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6bf5241c-b9fe-47ab-b323-9ed1930dd69f · outbound

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

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Sparse autoencoders find highly interpretable features in language models

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:50.993730Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0e64380a-0b3c-4000-8d3d-4939385c087f · outbound

This paper cites Eliciting secret knowledge from language models, 2025.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Eliciting secret knowledge from language models, 2025

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.146487Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 54730b82-5245-4b7e-bd5e-c4fd189d1119 · outbound

This paper cites Transcoders Find Interpretable LLM Feature Circuits.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Transcoders Find Interpretable LLM Feature Circuits

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.267585Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 86a42df9-f3e5-45d9-987c-d627751bd17a · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.365998Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 01b4dc7f-c90a-4f08-9947-af392d82d1a1 · outbound

This paper cites The Pile: An 800GB Dataset of Diverse Text for Language Modeling.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent The Pile: An 800GB Dataset of Diverse Text for Language Modeling

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.459639Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0c9c77ef-a42f-487f-84fb-9f7432ded972 · outbound

This paper cites Scaling and evaluating sparse autoencoders.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Scaling and evaluating sparse autoencoders

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.674341Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9afcc0d6-1975-4f39-9c4a-c8dd6b974cc7 · outbound

This paper cites SAGE: An agentic explainer framework for interpreting SAE features in language models.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent SAGE: An agentic explainer framework for interpreting SAE features in language models

Reference 12

Resolution: verified exact; doi, observed 2026-08-01T21:04:41.768292Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation bd08b618-42c6-4678-a0ba-604775c72ebf · outbound

This paper cites Current activation oracles are hard to use.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Current activation oracles are hard to use

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.887604Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 72329177-b199-4ebd-a356-53db9f1e9dea · outbound

This paper cites Diskann: Fast accurate billion-point nearest neighbor search on a single node.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Diskann: Fast accurate billion-point nearest neighbor search on a single node

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:51.946525Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 94c4b679-1eb6-48fd-b49b-8a2c50b873a3 · outbound

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

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Are sparse autoencoders useful? a case study in sparse probing

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.019579Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 962019f9-642b-403e-abb4-1b8514e2b249 · outbound

This paper cites Scaling Laws for Neural Language Models.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Scaling Laws for Neural Language Models

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.068818Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ade41b7f-e578-4028-bfbe-abfd3e7ee2e0 · outbound

This paper cites Activation oracles: Training and evaluating LLMs as general-purpose activation explainers.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Activation oracles: Training and evaluating LLMs as general-purpose activation explainers

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.134169Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c06181b7-bb80-49fb-bf7c-9ee1a759c4bf · outbound

This paper cites Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christopher Potts.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christopher Potts

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.171422Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9165bb8-c399-4174-b35d-66231afcdd6b · outbound

This paper cites SAEs are highly dataset dependent: a case study on the refusal direction.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent SAEs are highly dataset dependent: a case study on the refusal direction

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.259407Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a0506ede-addc-4ce2-8a63-4889a913a2e5 · outbound

This paper cites Sparse autoencoders do not find canonical units of analysis.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Sparse autoencoders do not find canonical units of analysis

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.344523Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f4674fb3-55c8-465d-a1cd-102d210b04f1 · outbound

This paper cites Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.501832Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 72eb814a-47cf-4fcb-9937-5cebae7139d2 · outbound

This paper cites Sparse crosscoders for cross-layer features and model diffing.Transformer Circuits Thread, 2024.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Sparse crosscoders for cross-layer features and model diffing.Transformer Circuits Thread, 2024

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.609252Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation de804e8d-bfff-44e3-9657-c7653eb08275 · outbound

This paper cites Linguistic regularities in continuous space word representations.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Linguistic regularities in continuous space word representations

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.725403Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e3c09eb6-dad5-46bf-821f-447b52c2ac9f · outbound

This paper cites Sparse Autoencoders for Hypothesis Generation.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Sparse Autoencoders for Hypothesis Generation

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:52.885406Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 24edcb8f-84ea-4815-8631-2b92f811cb28 · outbound

This paper cites All-but-the-top: Simple and effective postprocessing for word representations.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent All-but-the-top: Simple and effective postprocessing for word representations

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.084745Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3e46f9e5-bbe5-45be-8167-01b9e00f019c · outbound

This paper cites Latentqa: Teaching llms to decode activations into natural language.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Latentqa: Teaching llms to decode activations into natural language

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.191733Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a6303ff6-ef24-4008-b2f3-4b239bbaf0f7 · outbound

This paper cites Automatically inter- preting millions of features in large language models.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Automatically inter- preting millions of features in large language models

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.474769Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4cd54edd-cecb-43ab-af61-32eb1285c167 · outbound

This paper cites Topicgpt: A prompt-based topic modeling framework, 2024.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Topicgpt: A prompt-based topic modeling framework, 2024

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.604903Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9714d59b-c75b-4391-bfec-d09c4e429cd2 · outbound

This paper cites Svcca: Singular vector canonical correlation analysis for deep learning dynamics and interpretability.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Svcca: Singular vector canonical correlation analysis for deep learning dynamics and interpretability

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.762052Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1794cc39-dff3-4338-a6dc-9f05ac838eb5 · outbound

This paper cites Daniel Freeman, Theodore R.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Daniel Freeman, Theodore R

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.894752Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 67b37175-3b56-4910-984d-8f416ca019f6 · outbound

This paper cites AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.009938Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0c54125e-7c97-4cb5-8c6e-0129cd133422 · outbound

This paper cites ReAct: Synergizing reasoning and acting in language models.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent ReAct: Synergizing reasoning and acting in language models

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.066970Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3e26e2a5-b48e-403a-8cea-6547107104cb · outbound

This paper cites average of positives.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent average of positives

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.150881Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 532d191f-9aab-4c92-a5b1-5fd0f4acca3b · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.324830Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation abda397d-81ab-491b-8593-4e5dcb10bbcd · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.477338Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cdd929b4-a4b5-44c7-8565-2b4c294c5ece · outbound

This paper cites zero_vector.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent zero_vector

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.634866Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f28fba79-04e0-4ca8-a622-e72eb45df1b4 · outbound

This paper cites target_vector.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent target_vector

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.814870Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a08538b1-d39d-44ca-ae9a-707b7aac2b06 · outbound

This paper cites residual_after_<concep t>.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent residual_after_<concep t>

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:54.984575Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a797bef2-b6f5-47d2-8fee-d25a73ee1a5c · outbound

This paper cites coal_pc0.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent coal_pc0

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.142498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1e04671f-e494-42be-a6a9-d5a334de99b6 · outbound

This paper cites target_vector.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent target_vector

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.184833Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 90b5a10b-6f1e-446a-b917-feb258d8c05c · outbound

This paper cites Look for a coherent theme: a recurring topic, object, attribute, or domain that links most of the retrieved snippets.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Look for a coherent theme: a recurring topic, object, attribute, or domain that links most of the retrieved snippets

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.304955Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f9b542ff-0771-4a9b-8447-c902c33707f0 · outbound

This paper cites Ignore them.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Ignore them

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.634748Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 07cdbce7-39e5-4446-a841-c0003eb5a528 · outbound

This paper cites gender”, “sentiment.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent gender”, “sentiment

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.750130Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 695ac0ac-ba3f-4c2c-ac39-5f4d7a152fd8 · outbound

This paper cites positive-class.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent positive-class

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:55.918145Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5615ffd5-d350-4501-83bc-3f9f26882134 · outbound

This paper cites Use simple stable names likepos_0, pos_1, ...andneg_0, neg_1,.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Use simple stable names likepos_0, pos_1, ...andneg_0, neg_1,

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.094829Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1bfe1313-d36b-4b13-9082-72dd0994d0a8 · outbound

This paper cites <concept_name>_direction.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent <concept_name>_direction

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.192082Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8a0ea669-fbcf-4a9c-bbce-63b95f0e55e2 · outbound

This paper cites target_vector.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent target_vector

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.366713Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5b589d74-5c94-425e-adf0-532570092df7 · outbound

This paper cites gender_dir ection.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent gender_dir ection

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.511250Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 57bf2ecd-bb80-495b-8073-6c4b8f020acc · outbound

This paper cites legal language.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent legal language

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.624745Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 730b701e-3383-45b6-aa6c-a85c8d7dbdf9 · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.751608Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fae3cea8-0973-49af-9eb5-73d3be4c350a · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.848084Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a939941d-4950-4c01-ac14-5115ff1730ea · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:56.996961Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 95e36f9c-7a34-4b1a-bc9a-e1ddf41fb85e · outbound

This paper cites we were unable to find the license for the dataset we used.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent we were unable to find the license for the dataset we used

Reference 54

Resolution: malformed identifier; no resolver link, observed 2026-08-01T20:58:57.078133Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2ebccd26-3947-4038-bc2d-252b43418832 · outbound

This paper cites • Depending on the country in which research is conducted, IRB approval (or equivalent) may be required for any human subjects research.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent • Depending on the country in which research is conducted, IRB approval (or equivalent) may be required for any human subjects research

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:57.131962Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a8795be1-04e3-415a-9ebf-3d628077af3b · outbound

This paper cites an unresolved cited work.

Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent Unresolved cited work

Reference 2026

Resolution: unresolved; no resolver link, observed 2026-08-01T20:58:53.333953Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers