31 citations · 180 across the 32 of their papers we have counts for
36 papers
CT: Counterfactual Causal Reasoning for Sentiment Shifts in Social-Media Conversation Trees
S M Rafiuddin, Atriya Sen
Sentiment in social-media threads does not only vary across posts; it shifts as users react to claims, corrections, evidence, and hostility within a branching reply tree. We study…
Rethinking Thinking Tokens: LLMs as Improvement Operators
Lovish Madaan, Aniket Didolkar, Suchin Gururangan +6
Reasoning training incentivizes LLMs to produce long chains of thought (long CoT), which among other things, allows them to explore solution strategies with self-checking. This res…
Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
Dulhan Jayalath, Shashwat Goel, Thomas Foster +5
Where do learning signals come from when there is no ground truth in post-training? We show that inference compute itself can serve as supervision. By generating parallel rollouts…
From KMMLU-Redux to KMMLU-Pro: A Professional Korean Benchmark Suite for LLM Evaluation
Seokhee Hong, Sunkyoung Kim, Guijin Son +3
The development of Large Language Models (LLMs) requires robust benchmarks that encompass not only academic domains but also industrial fields to effectively evaluate their applica…
Balancing Knowledge Delivery and Emotional Comfort in Healthcare Conversational Systems
Shang-Chi Tsai, Yun-Nung Chen
With the advancement of large language models, many dialogue systems are now capable of providing reasonable and informative responses to patients' medical conditions. However, whe…
Dialect Normalization using Large Language Models and Morphological Rules
Antonios Dimakis, John Pavlopoulos, Antonios Anastasopoulos
Natural language understanding systems struggle with low-resource languages, including many dialects of high-resource ones. Dialect-to-standard normalization attempts to tackle thi…
Jailbreak Distillation: Renewable Safety Benchmarking
Jingyu Zhang, Ahmed Elgohary, Xiawei Wang +5
Large language models (LLMs) are rapidly deployed in critical applications, raising urgent needs for robust safety benchmarking. We propose Jailbreak Distillation (JBDistill), a no…
Diversity-driven Data Selection for Language Model Tuning through Sparse Autoencoder
Xianjun Yang, Shaoliang Nie, Lijuan Liu +5
Instruction tuning data are often quantity-saturated due to the large volume of data collection and fast model iteration, leaving data selection important but underexplored. Existi…
BTS: Harmonizing Specialized Experts into a Generalist LLM
Qizhen Zhang, Prajjwal Bhargava, Chloe Bi +9
We present Branch-Train-Stitch (BTS), an efficient and flexible training algorithm for combining independently trained large language model (LLM) experts into a single, capable gen…
Self-Generated Critiques Boost Reward Modeling for Language Models
Yue Yu, Zhengxing Chen, Aston Zhang +10
Reward modeling is crucial for aligning large language models (LLMs) with human preferences, especially in reinforcement learning from human feedback (RLHF). However, current rewar…
The Llama 3 Herd of Models
Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri +556
Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models th…
DataComp-LM: In search of the next generation of training sets for language models
Jeffrey Li, Alex Fang, Georgios Smyrnis +56
We introduce DataComp for Language Models (DCLM), a testbed for controlled dataset experiments with the goal of improving language models. As part of DCLM, we provide a standardize…
Language models scale reliably with over-training and on downstream tasks
Samir Yitzhak Gadre, Georgios Smyrnis, Vaishaal Shankar +22
Scaling laws are useful guides for derisking expensive training runs, as they predict performance of large models using cheaper, small-scale experiments. However, there remain gaps…
LESS: Selecting Influential Data for Targeted Instruction Tuning
Mengzhou Xia, Sadhika Malladi, Suchin Gururangan +2
Instruction tuning has unlocked powerful capabilities in large language models (LLMs), effectively using combined datasets to develop generalpurpose chatbots. However, real-world a…
Breaking the Curse of Multilinguality with Cross-lingual Expert Language Models
Terra Blevins, Tomasz Limisiewicz, Suchin Gururangan +4
Despite their popularity in non-English NLP, multilingual language models often underperform monolingual ones due to inter-language competition for model parameters. We propose Cro…
AboutMe: Using Self-Descriptions in Webpages to Document the Effects of English Pretraining Data Filters
Li Lucy, Suchin Gururangan, Luca Soldaini +4
Large language models' (LLMs) abilities are drawn from their pretraining data, and model development begins with data curation. However, decisions around what data is retained or r…
Time is Encoded in the Weights of Finetuned Language Models
Kai Nylund, Suchin Gururangan, Noah A. Smith
We present time vectors, a simple tool to customize language models to new time periods. Time vectors are created by finetuning a language model on data from a single time (e.g., a…
SILO Language Models: Isolating Legal Risk In a Nonparametric Datastore
Sewon Min, Suchin Gururangan, Eric Wallace +4
The legality of training language models (LMs) on copyrighted or otherwise restricted data is under intense debate. However, as we show, model performance significantly degrades if…
Information Flow Control in Machine Learning through Modular Model Architecture
Trishita Tiwari, Suchin Gururangan, Chuan Guo +7
In today's machine learning (ML) models, any part of the training data can affect the model output. This lack of control for information flow from training data to model output is…
Scaling Expert Language Models with Unsupervised Domain Discovery
Suchin Gururangan, Margaret Li, Mike Lewis +4
Large language models are typically trained densely: all parameters are updated with respect to all inputs. This requires synchronization of billions of parameters across thousands…
Editing Models with Task Arithmetic
Gabriel Ilharco, Marco Tulio Ribeiro, Mitchell Wortsman +4
Changing how pre-trained models behave -- e.g., improving their performance on a downstream task or mitigating biases learned during pre-training -- is a common practice when devel…
lo-fi: distributed fine-tuning without communication
Mitchell Wortsman, Suchin Gururangan, Shen Li +4
When fine-tuning large neural networks, it is common to use multiple nodes and to communicate gradients at each optimization step. By contrast, we investigate completely local fine…
M2D2: A Massively Multi-domain Language Modeling Dataset
Machel Reid, Victor Zhong, Suchin Gururangan +1
We present M2D2, a fine-grained, massively multi-domain corpus for studying domain adaptation in language models (LMs). M2D2 consists of 8.5B tokens and spans 145 domains extracted…
Branch-Train-Merge: Embarrassingly Parallel Training of Expert Language Models
Margaret Li, Suchin Gururangan, Tim Dettmers +4
We present Branch-Train-Merge (BTM), a communication-efficient algorithm for embarrassingly parallel training of large language models (LLMs). We show it is possible to independent…