Paper Citation Record · LEDGER

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge

As of 20 August 2026, this Paper Citation Record lists 32 of 32 outbound references and 2 inbound Pith citation observations for arXiv:2505.05946.

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

pith.paper-citation-record.v1
2505.05946 v2

Coverage vector

measured 32 of 32 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-15T22:55:45.986790Z

measured 34 of 34 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 2 of 2 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-06-27T05:02:18.347642Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-07-03T16:48:39.928976Z

Reference resolution

32 of 32 outbound references displayed

verified exact2
verified fuzzy19
unresolved11
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation e5dc1b4d-340e-4485-b6c0-eeaa64e97972 · outbound

This paper cites Attention is all you need.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Attention is all you need

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.525323Z

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 82b5ce1d-8cea-4db4-a9a3-4c6145c21750 · outbound

This paper cites Bring Your Own Knowledge: A Survey of Methods for LLM Knowledge Expansion, 2025.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Bring Your Own Knowledge: A Survey of Methods for LLM Knowledge Expansion, 2025

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.511054Z

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 d0acec44-9979-4bde-9d5f-3b714c70d28c · outbound

This paper cites Continual Learning of Large Language Models: A Comprehensive Survey.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Continual Learning of Large Language Models: A Comprehensive Survey

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.849387Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ba9bc21-6ca5-4d23-a298-34ad7dffefc3 · outbound

This paper cites The MIT Press, Cambridge, 1965.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge The MIT Press, Cambridge, 1965

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.496916Z

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 50ccf73e-11ea-4c52-99bf-90a2baf301be · outbound

This paper cites Overcoming Catastrophic Forgetting in Neural Networks.Proceedings of the national academy of sciences, 114(13):3521–3526, 2017.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Overcoming Catastrophic Forgetting in Neural Networks.Proceedings of the national academy of sciences, 114(13):3521–3526, 2017

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.481460Z

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 11daf383-8223-4096-9b71-6234ddcdae9d · outbound

This paper cites Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.864261Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation afbcb2d5-bdaa-489d-8914-ea1258c57fc7 · outbound

This paper cites Fine-tuned Language Models are Continual Learners.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Fine-tuned Language Models are Continual Learners

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.869630Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c87e7829-d4d4-42bb-b8a4-3c06c09805b7 · outbound

This paper cites Simple and Scalable Strategies to Continually Pre-train Large Language Models.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Simple and Scalable Strategies to Continually Pre-train Large Language Models

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.874320Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b1715bdd-a4a9-4a41-a736-060a0707cba3 · outbound

This paper cites LAMOL: LAnguage MOdeling for Lifelong Language Learning.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge LAMOL: LAnguage MOdeling for Lifelong Language Learning

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.879138Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4aac1fb0-96ed-434c-b2f5-c159df85c7b9 · outbound

This paper cites Learning Without Forgetting.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Learning Without Forgetting

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.465648Z

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 0a18f165-9e35-4ddc-8177-cd0202eb1982 · outbound

This paper cites Learning to solve NLP tasks in an incremental number of languages.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Learning to solve NLP tasks in an incremental number of languages

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.449852Z

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 b9b2c15b-b3cc-41fd-80c1-366af21d704d · outbound

This paper cites LoRA: Low-rank adaptation of large language models.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge LoRA: Low-rank adaptation of large language models

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.434986Z

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 be4ad56f-ed2c-4785-a315-76b278609598 · outbound

This paper cites CURLoRA: Stable LLM Continual Fine-Tuning and Catastrophic Forgetting Mitigation, 2024.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge CURLoRA: Stable LLM Continual Fine-Tuning and Catastrophic Forgetting Mitigation, 2024

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.420189Z

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 b4a70fec-bc2a-4132-b0e8-851e5b0845a9 · outbound

This paper cites Language models meet world models: Embodied experiences enhance language models.Advances in Neural Information Process- ing Systems, 36:75392–75412, 2023.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Language models meet world models: Embodied experiences enhance language models.Advances in Neural Information Process- ing Systems, 36:75392–75412, 2023

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.405138Z

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 f45a1dd0-2145-4bad-8713-b9b723a27b0c · outbound

This paper cites Overcoming Catastrophic Forgetting in Massively Multilingual Continual Learning.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Overcoming Catastrophic Forgetting in Massively Multilingual Continual Learning

Reference 15

Resolution: verified exact; local_arxiv, observed 2026-08-15T22:55:46.152730Z

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 35cf8420-a712-4207-b0fc-bc9aad3f22b2 · outbound

This paper cites Unifying Importance Based Regularisation Methods for Continual Learning.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Unifying Importance Based Regularisation Methods for Continual Learning

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.390008Z

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 4b7969c5-532f-4591-9a44-cb181917839f · outbound

This paper cites van de Ven.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge van de Ven

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.374637Z

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 f642ff60-0197-4dd5-b20a-0de9ba0a6605 · outbound

This paper cites Examining Forgetting in Continual Pre-training of Aligned Large Language Models.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Examining Forgetting in Continual Pre-training of Aligned Large Language Models

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.920900Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dd73bfee-9bfd-4836-b56f-55341cbfd7aa · outbound

This paper cites Gemma 2: Improving Open Language Models at a Practical Size.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Gemma 2: Improving Open Language Models at a Practical Size

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.925472Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f8619775-8d88-48de-aede-55cc71934158 · outbound

This paper cites CulturaX: A Cleaned, Enormous, and Multilingual Dataset for Large Language Models in 167 Languages.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge CulturaX: A Cleaned, Enormous, and Multilingual Dataset for Large Language Models in 167 Languages

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.930374Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1010db17-f03f-4b60-b256-d71eb4ba6c20 · outbound

This paper cites Open Llama2 Model for the Lithuanian Language.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Open Llama2 Model for the Lithuanian Language

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.935063Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 46c2bc39-d873-4557-9116-1bbd1280a35d · outbound

This paper cites Open Llama2 Models for the Lithuanian Language.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Open Llama2 Models for the Lithuanian Language

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.358723Z

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 55dfaadb-56fb-4c26-8bdd-f836e10d5259 · outbound

This paper cites Localizing AI: Evaluating Open-Weight Language Models for Languages of Baltic States.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Localizing AI: Evaluating Open-Weight Language Models for Languages of Baltic States

Reference 23

Resolution: verified exact; local_arxiv, observed 2026-08-15T22:55:46.066445Z

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 1c5527ce-923f-40fb-82bd-cdc2204e58a6 · outbound

This paper cites Align- ing AI With Shared Human Values.Proceedings of the International Conference on Learning Representations (ICLR), 2021.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Align- ing AI With Shared Human Values.Proceedings of the International Conference on Learning Representations (ICLR), 2021

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.343551Z

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 fbb2e76e-b153-49db-be2d-cb1cb68e88d6 · outbound

This paper cites Measuring Massive Multitask Language Understanding.Proceedings of the International Conference on Learn- ing Representations (ICLR), 2021.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Measuring Massive Multitask Language Understanding.Proceedings of the International Conference on Learn- ing Representations (ICLR), 2021

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.326789Z

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 7904a22f-f143-410a-817a-d1ecd5b43ba1 · outbound

This paper cites The Belebele Benchmark: a Parallel Reading Comprehension Dataset in 122 Language Variants.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge The Belebele Benchmark: a Parallel Reading Comprehension Dataset in 122 Language Variants

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.310535Z

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 3a1c596b-a8d1-4605-ab35-95297da8aa0b · outbound

This paper cites Training Verifiers to Solve Math Word Problems.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Training Verifiers to Solve Math Word Problems

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.962563Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2607462f-48a4-4ab3-935d-56a9f8e7245b · outbound

This paper cites HellaSwag: Can a Machine Re- ally Finish Your Sentence? InProceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge HellaSwag: Can a Machine Re- ally Finish Your Sentence? InProceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.294473Z

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 f34f9659-586c-4152-97fc-962a851e78f1 · outbound

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

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-15T22:55:45.972656Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2e8c0c0b-6045-4134-8be8-5dacd86a329c · outbound

This paper cites TruthfulQA: Measuring How Models Mimic Human Falsehoods, 2021.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge TruthfulQA: Measuring How Models Mimic Human Falsehoods, 2021

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.278949Z

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 2452316c-e0b8-44e1-8dbd-b9851b774e45 · outbound

This paper cites WinoGrande: An Adversarial Wino- grad Schema Challenge at Scale.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge WinoGrande: An Adversarial Wino- grad Schema Challenge at Scale

Reference 31

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.263222Z

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 0f505a17-c735-41c2-9ee5-498617789b14 · outbound

This paper cites Fine-tuned.

Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge Fine-tuned

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-15T22:55:46.247316Z

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

Observation ceada456-ab38-4e67-9a76-b76b836438e7 · inbound

TFGN: Task-Free, Replay-Free Continual Pre-Training Without Catastrophic Forgetting at LLM Scale cites this paper.

TFGN: Task-Free, Replay-Free Continual Pre-Training Without Catastrophic Forgetting at LLM Scale Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge

Reference 46

Resolution: verified exact; arxiv_id, observed 2026-05-19T17:07:41.541141Z

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 cdd7dc52-bb2c-4d2f-8962-77f272e28d1d · inbound

LLM Evolution as an Industry-Scale Ecosystem: A Lifecycle Perspective on Continual Learning cites this paper.

LLM Evolution as an Industry-Scale Ecosystem: A Lifecycle Perspective on Continual Learning Full-Parameter Continual Pretraining of Gemma2: Insights into Fluency and Domain Knowledge

Reference 150

Resolution: verified exact; arxiv_id, observed 2026-07-03T16:48:39.930294Z

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.