Papers with large models
Through the Valley: Path to Effective Long CoT Training for Small Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Long chain-of-thought (CoT) supervision is effective for large language models . but small models trained on limited long CoT data experience performance degradation . |
| Approach: | They identify a phenomenon called Long CoT Degradation in small language models . long CoT data can be used to generate long chain-of-thought (CoT) responses . |
| Outcome: | The results show that models trained on 8k long CoT examples lose up to 75% of their original performance before fine-tuning. |
GrIPS: Gradient-free, Edit-based Instruction Search for Prompting Large Language Models (2023.eacl-main)
Copied to clipboard
| Challenge: | Recent work aimed to improve task performance of large language models by rewriting or tuning them manually, but manual rewrite is time-consuming and requires subjective interpretation. |
| Approach: | They propose a gradient-free, edit-based search approach for improving task instructions for large language models. |
| Outcome: | The proposed approach outperforms manual rewriting and purely example-based prompts while allowing for API-based tuning. |
Dynamically Adjusting Transformer Batch Size by Monitoring Gradient Direction Change (2020.acl-main)
Copied to clipboard
| Challenge: | Compared to previous studies, the performance of neural models is likely to be affected by the choice of hyper-parameters. |
| Approach: | They propose to automatically and dynamically determine batch sizes by accumulating gradients of mini-batches and performing an optimization step at just the time when the direction of gradients starts to fluctuate. |
| Outcome: | The proposed approach improves the Transformer model with a fixed 25k batch size by +0.73 and +0.82 BLEU respectively. |
Enhancing Translation Ability of Large Language Models by Leveraging Task-Related Layers (2024.lrec-main)
Copied to clipboard
| Challenge: | Experimental validation shows that adjusting task-related layers significantly improves performance on translation tasks while maintaining stability and accuracy on other tasks. |
| Approach: | They propose to adjust task-related layers in large models to better harness their machine translation capabilities by revealing the structure and characteristics of attention weights through singular value decomposition. |
| Outcome: | The proposed method reduces computational resource consumption and catastrophic forgetting while maintaining stability and accuracy on other tasks. |
Speculative Safety-Aware Decoding (2025.emnlp-main)
Copied to clipboard
| Challenge: | Speculative Safety-Aware Decoding (SSD) equips large language models with desired safety property while accelerating inference. |
| Approach: | They propose a lightweight decoding-time approach that equips large models with the desired safety property while accelerating inference. |
| Outcome: | Experimental results show that a small language model has the desired safety property while accelerating inference. |
BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting (2023.acl-long)
Copied to clipboard
Zheng Xin Yong, Hailey Schoelkopf, Niklas Muennighoff, Alham Fikri Aji, David Ifeoluwa Adelani, Khalid Almubarak, M Saiful Bari, Lintang Sutawika, Jungo Kasai, Ahmed Baruwa, Genta Winata, Stella Biderman, Edward Raff, Dragomir Radev, Vassilina Nikoulina
| Challenge: | Existing language adaptation strategies for multilingual models are limited to 46 languages . a new language is added to the model to improve zero-shot prompting performance . |
| Approach: | They apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. |
| Outcome: | The proposed model can be extended to other languages without incurring prohibitively large costs. |