Paper Citation Record · LEDGER

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories

As of 21 August 2026, this Paper Citation Record lists 18 of 18 outbound references and 1 inbound Pith citation observation for arXiv:2606.18663.

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

pith.paper-citation-record.v1
2606.18663 v1

Coverage vector

measured 18 of 18 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-06-26T21:01:52.850037Z

measured 19 of 19 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-20T06:33:59.587034+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-07-31T12:45:04.860018Z

measured 0 of 1 external citation measurements

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

Source: cited_works

Reference resolution

18 of 18 outbound references displayed

verified exact1
verified fuzzy0
unresolved5
parse uncertain0
malformed identifier0
metadata mismatch12

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation c0dcb010-0124-49c0-b659-74bd88b45218 · outbound

This paper cites Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge

Reference 1

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.553085Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 0e1ab5ef-6b86-46ed-a171-5be072b032b1 · outbound

This paper cites Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining

Reference 2

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.556230Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 4019f4de-d7b8-49c4-a154-e47b2509ba3b · outbound

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

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories The Pile: An 800GB Dataset of Diverse Text for Language Modeling

Reference 3

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.552904Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d94a8064-3839-4ad5-b279-7b3f0505cdfe · outbound

This paper cites Scaling Laws for Neural Language Models.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Scaling Laws for Neural Language Models

Reference 4

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.546858Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation aefbebaa-8c51-44ec-9920-27463424066b · outbound

This paper cites an unresolved cited work.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Unresolved cited work

Reference 5

Resolution: unresolved; no resolver link, observed 2026-06-26T21:01:52.850037Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 22294470-aed9-49ba-9166-c8b75cbd77e1 · outbound

This paper cites In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 785– 794, Copenhagen, Denmark.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 785– 794, Copenhagen, Denmark

Reference 6

Resolution: unresolved; no resolver link, observed 2026-06-26T21:01:52.850037Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 699bda87-0340-45ae-93ba-1481266202f4 · outbound

This paper cites Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training

Reference 7

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.537860Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 489123d4-9d43-41e4-af03-6104752a8a4d · outbound

This paper cites LogiQA: A Challenge Dataset for Machine Reading Comprehension with Logical Reasoning.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories LogiQA: A Challenge Dataset for Machine Reading Comprehension with Logical Reasoning

Reference 8

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.537323Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation b9e68ac3-901a-49f9-9706-64114889e587 · outbound

This paper cites Actor-critic based online data mixing for language model pre-training.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Actor-critic based online data mixing for language model pre-training

Reference 9

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.535109Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation ecfa4696-2537-4850-8d61-a12e22caf124 · outbound

This paper cites Mid-training of large language models: A survey.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Mid-training of large language models: A survey

Reference 10

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.559034Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation c80f5aa4-5e42-4fa7-929a-96ba99600a4c · outbound

This paper cites an unresolved cited work.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Unresolved cited work

Reference 11

Resolution: unresolved; no resolver link, observed 2026-06-26T21:01:52.850037Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 849c506c-7a93-4ba5-9b91-5648810c9074 · outbound

This paper cites WinoGrande: An Adversarial Winograd Schema Challenge at Scale.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories WinoGrande: An Adversarial Winograd Schema Challenge at Scale

Reference 12

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.528125Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8659ef9b-b49a-45a2-836c-a48f5c29553e · outbound

This paper cites InProceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 353–355, Brussels, Belgium.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories InProceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 353–355, Brussels, Belgium

Reference 13

Resolution: unresolved; no resolver link, observed 2026-06-26T21:01:52.850037Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f8b5a224-a9e0-4bf7-baf7-91a14c9c785d · outbound

This paper cites Mergemix: Optimizing mid-training data mixtures via learnable model merging.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Mergemix: Optimizing mid-training data mixtures via learnable model merging

Reference 14

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.524351Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation e0262ce9-b9e8-49f8-b9df-7a43575e99d6 · outbound

This paper cites TiKMiX: Take Data Influence into Dynamic Mixture for Language Model Pre-training.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories TiKMiX: Take Data Influence into Dynamic Mixture for Language Model Pre-training

Reference 15

Resolution: metadata mismatch; arxiv_id, observed 2026-07-04T00:39:17.540921Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 2920864c-b146-4915-bb6e-d750804183b7 · outbound

This paper cites TinyLlama: An Open-Source Small Language Model.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories TinyLlama: An Open-Source Small Language Model

Reference 16

Resolution: metadata mismatch; local_arxiv, observed 2026-07-04T00:39:17.543672Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 9d36d063-dda9-4aad-9111-0812867c0288 · outbound

This paper cites Proxy models are trained on 1 H800 GPU for 1,000 steps (1M tokens per step) and the target model is trained on 8 GPUs for 25,000 steps, totaling 25B tokens.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories Proxy models are trained on 1 H800 GPU for 1,000 steps (1M tokens per step) and the target model is trained on 8 GPUs for 25,000 steps, totaling 25B tokens

Reference 17

Resolution: unresolved; no resolver link, observed 2026-06-26T21:01:52.850037Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ff8a444e-eb64-4f55-b07f-0a8e434be125 · outbound

This paper cites We use the Pile-CC validation loss as the target predicted loss.

RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories We use the Pile-CC validation loss as the target predicted loss

Reference 18

Resolution: verified exact; arxiv_id, observed 2026-07-04T00:39:17.521440Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Pith citing papers

Observation 3a1f65f6-93cd-4aeb-9728-8c0c636b21e0 · inbound

DecoupleMix: Decoupled Ratio Search and Convex Allocation for Scalable VLM Data Recipes cites this paper.

DecoupleMix: Decoupled Ratio Search and Convex Allocation for Scalable VLM Data Recipes RegMix-D: Dynamic Data Mixing via Proxy Training Trajectories

Reference 26

Resolution: unresolved; no resolver link, observed 2026-07-31T12:45:04.860018Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.