Papers by Fartash Faghri
MUSCLE: A Model Update Strategy for Compatible LLM Evolution (2024.findings-emnlp)
Copied to clipboard
Jessica Echterhoff, Fartash Faghri, Raviteja Vemulapalli, Ting-Yao Hu, Chun-Liang Li, Oncel Tuzel, Hadi Pouransari
| Challenge: | Large Language Models (LLMs) are regularly updated to enhance performance, typically through changes in data or architecture. |
| Approach: | They propose a method to minimize the extent of instance regression in model updates by training a compatibility adapter that can enhance task fine-tuned language models. |
| Outcome: | The proposed approach reduces regressions by up to 40% when updating LLMs to newer versions while maintaining overall performance gains. |
TiC-LM: A Web-Scale Benchmark for Time-Continual LLM Pretraining (2025.acl-long)
Copied to clipboard
Jeffrey Li, Mohammadreza Armandpour, Seyed Iman Mirzadeh, Sachin Mehta, Vaishaal Shankar, Raviteja Vemulapalli, Samy Bengio, Oncel Tuzel, Mehrdad Farajtabar, Hadi Pouransari, Fartash Faghri
| Challenge: | Large language models (LLMs) trained on historical web data inevitably become outdated. |
| Approach: | They propose a web-scale dataset for time-continual pretraining of LLMs derived from 114 dumps of Common Crawl (CC) they also design time-stratified evaluations to assess how well various continual learning methods adapt to new data while retaining past knowledge. |
| Outcome: | The proposed benchmarks show that autoregressive meta-schedules combined with a fixed-ratio replay of older data can achieve comparable held-out loss to re-training from scratch, while requiring significantly less computation (2.6x). |
Beyond a Single Extractor: Re-thinking HTML-to-Text Extraction for LLM Pre-training (2026.findings-eacl)
Copied to clipboard
Jeffrey Li, Joshua P Gardner, Doug Kang, Fangping Shi, Karanjeet Singh, Chun-Liang Li, Herumb Shandilya, David Leo Wright Hall, Oncel Tuzel, Percy Liang, Ludwig Schmidt, Hadi Pouransari, Fartash Faghri
| Challenge: | Existing open-source datasets predominantly apply a single fixed extractor to all webpages. |
| Approach: | They propose to take a Union over different extractors to improve model performance . they show that extractor choice can significantly impact downstream task performance based on content type . |
| Outcome: | The proposed approach can increase the token yield of DCLM-Baseline by 71% while maintaining benchmark performance. |