Papers with continual training
LADM: Long-context Training Data Selection with Attention-based Dependency Measurement for LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Long-context modeling has drawn more attention in the area of Large Language Models (LLMs). |
| Approach: | They propose a Long-context data selection framework with Attention-based Dependency Measurement which can efficiently identify high-quality long-contrast data from a large-scale, multi-domain pre-training corpus. |
| Outcome: | The proposed framework significantly boosts the performance of LLMs on multiple long-context tasks with only 1B tokens for continual training. |
Efficient Long-Range Transformers: You Need to Attend More, but Not Necessarily at Every Layer (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Pretrained transformer models have demonstrated remarkable performance across various natural language processing tasks. |
| Approach: | They propose a transformer variant with mixed attention spans that leverages the attention mechanism to capture long- and short-range dependencies in the sequence. |
| Outcome: | The proposed model can achieve competitive performance to models with full attention while reducing computational cost (75%) |
Beyond Random Sampling: Efficient Language Model Pretraining via Curriculum Learning (2026.eacl-long)
Copied to clipboard
| Challenge: | Curriculum learning has improved efficiency across machine learning domains, but remains underexplored for language model pretraining. |
| Approach: | They present a systematic investigation of curriculum learning in LLM pretraining . they use vanilla curriculum learning, pacing-based sampling, and interleaved curricula . |
| Outcome: | The proposed framework accelerates convergence in early and mid-training phases, reducing training steps by 18-45% to reach baseline performance. |
Optimizing LLMs for Italian: Reducing Token Fertility and Enhancing Efficiency Through Vocabulary Adaptation (2025.findings-naacl)
Copied to clipboard
Luca Moroni, Giovanni Puccetti, Pere-Lluís Huguet Cabot, Andrei Stefan Bejgu, Alessio Miaschi, Edoardo Barba, Felice Dell’Orletta, Andrea Esuli, Roberto Navigli
| Challenge: | Pretrained Large Language Models (LLMs) are mainly designed for the English language, but are not optimized for non-English languages due to language contamination or multilingual pretraining data. |
| Approach: | They propose a method that leverages neural mapping for vocabulary substitution to optimize LLMs for the Italian language. |
| Outcome: | The proposed method reduces token fertility by 25% and improves grounded alignment strategies. |
Investigating Catastrophic Forgetting During Continual Training for Neural Machine Translation (2020.coling-main)
Copied to clipboard
| Challenge: | Neural machine translation models suffer from catastrophic forgetting during continual training . models tend to overfit to frequent observations in the in-domain data but forget previously learned knowledge. |
| Approach: | They investigated the causes of catastrophic forgetting in NMT models by examining their parameters and modules. |
| Outcome: | The proposed model forgets previously learned knowledge and swings to fit new data . the results show that some parameters are important for both the general-domain and in-domain translation and the great change of them during continual training brings about the performance decline in general- domain. |
Improving Temporal Generalization of Pre-trained Language Models with Lexical Semantic Change (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to improve neural language models perform poorly on emerging data. |
| Approach: | They propose a lexical-level masking strategy to post-train a neural language model using static data from past years. |
| Outcome: | The proposed method outperforms existing methods on two pre-trained language models, two classification tasks, and four benchmark datasets. |
MixLLM: Dynamic Routing in Mixed Large Language Models (2025.naacl-long)
Copied to clipboard
Xinyuan Wang, Yanchi Liu, Wei Cheng, Xujiang Zhao, Zhengzhang Chen, Wenchao Yu, Yanjie Fu, Haifeng Chen
| Challenge: | Large Language Models (LLMs) exhibit potential artificial generic intelligence, however, their usage is costly with high response latency. |
| Approach: | They develop a dynamic contextual-bandit-based routing system for query-LLM assignment that leverages query tags to enhance query embeddings. |
| Outcome: | The proposed model maximizes response quality and minimizes cost and latency. |
ViLL-E: Video LLM Embeddings for Retrieval (2026.acl-long)
Copied to clipboard
| Challenge: | Video Large Language Models excel at video understanding tasks where outputs are textual . however, they underperform specialized embedding-based models in Retrieval tasks . |
| Approach: | They propose a video-LLM-based model with an embedding generation mechanism that allows the model to "think longer" for complex videos and stop early for easy ones. |
| Outcome: | The proposed model outperforms specialized embedding-based models in video understanding tasks while remaining competitive on VideoQA tasks. |