Kashish Mittal

Mohi Reza, Nathan Laundry, Ilya Musabirov et al.

Exploring alternative ideas by rewriting text is integral to the writing process. State-of-the-art Large Language Models (LLMs) can simplify writing variation generation. However, current interfaces pose challenges for simultaneous consideration of multiple variations: creating new variations without overwriting text can be difficult, and pasting them sequentially can clutter documents, increasing workload and disrupting writers' flow. To tackle this, we present ABScribe, an interface that supports rapid, yet visually structured, exploration and organization of writing variations in human-AI co-writing tasks. With ABScribe, users can swiftly modify variations using LLM prompts, which are auto-converted into reusable buttons. Variations are stored adjacently within text fields for rapid in-place comparisons using mouse-over interactions on a popup toolbar. Our user study with 12 writers shows that ABScribe significantly reduces task workload (d = 1.20, p < 0.001), enhances user perceptions of the revision process (d = 2.41, p < 0.001) compared to a popular baseline workflow, and provides insights into how writers explore variations using LLMs.

Tanmay Rajore, Nishanth Chandran, Sunayana Sitaram et al.

Benchmarking is the de-facto standard for evaluating LLMs, due to its speed, replicability and low cost. However, recent work has pointed out that the majority of the open source benchmarks available today have been contaminated or leaked into LLMs, meaning that LLMs have access to test data during pretraining and/or fine-tuning. This raises serious concerns about the validity of benchmarking studies conducted so far and the future of evaluation using benchmarks. To solve this problem, we propose Private Benchmarking, a solution where test datasets are kept private and models are evaluated without revealing the test data to the model. We describe various scenarios (depending on the trust placed on model owners or dataset owners), and present solutions to avoid data contamination using private benchmarking. For scenarios where the model weights need to be kept private, we describe solutions from confidential computing and cryptography that can aid in private benchmarking. We build an end-to-end system, TRUCE, that enables such private benchmarking showing that the overheads introduced to protect models and benchmark are negligible (in the case of confidential computing) and tractable (when cryptographic security is required). Finally, we also discuss solutions to the problem of benchmark dataset auditing, to ensure that private benchmarks are of sufficiently high quality.

Kashish Mittal, Di Yu, Roozbeh Ketabi et al.

Training massive-scale deep learning models on datasets spanning tens of terabytes presents critical challenges in hardware utilization and training reproducibility. In this paper, we identify and resolve profound data-loading bottlenecks within distributed GPU training pipelines using the Petastorm data loader and Apache Parquet datasets. Through systematic profiling, we demonstrate that network I/O and CPU-bound data transformations (e.g., PyArrow to NumPy) constrain GPU utilization to as low as 10-15%. To address this, we propose an optimized architecture that features push-down worker-level transformations coupled with local-disk caching via Fanout-Cache, minimizing redundant I/O and CPU overhead across training epochs. Furthermore, we eliminate race conditions in multi-worker shared queues by implementing dedicated round-robin ventilator and result queues, alongside modernized RNG handling, achieving strict deterministic data loading. Our optimizations yield a 6x speedup, reducing end-to-end training time from 22 hours to 3 hours, increasing GPU utilization to over 60%, and drastically reducing run-to-run variance, enabling robust, high-throughput, and reproducible large-scale model training.

Harsh Shah, Kashish Mittal, Ajit Rajwade

This work presents an adaptive group testing framework for the range-based high dimensional near neighbor search problem. Our method efficiently marks each item in a database as neighbor or non-neighbor of a query point, based on a cosine distance threshold without exhaustive search. Like other methods for large scale retrieval, our approach exploits the assumption that most of the items in the database are unrelated to the query. However, it does not assume a large difference between the cosine similarity of the query vector with the least related neighbor and that with the least unrelated non-neighbor. Following a multi-stage adaptive group testing algorithm based on binary splitting, we divide the set of items to be searched into half at each step, and perform dot product tests on smaller and smaller subsets, many of which we are able to prune away. We show that, using softmax-based features, our method achieves a more than ten-fold speed-up over exhaustive search with no loss of accuracy, on a variety of large datasets. Based on empirically verified models for the distribution of cosine distances, we present a theoretical analysis of the expected number of distance computations per query and the probability that a pool will be pruned. Our method has the following features: (i) It implicitly exploits useful distributional properties of cosine distances unlike other methods; (ii) All required data structures are created purely offline; (iii) It does not impose any strong assumptions on the number of true near neighbors; (iv) It is adaptable to streaming settings where new vectors are dynamically added to the database; and (v) It does not require any parameter tuning. The high recall of our technique makes it particularly suited to plagiarism detection scenarios where it is important to report every database item that is sufficiently similar item to the query.