Constantin Adam

Nelson Bore, Pritish Mishra, Constantin Adam et al.

Fuzzy deduplication is key to constructing large language model training corpora. However, classic Locality-Sensitive Hashing pipelines scale poorly as corpora grow and are ill-suited to continuous ingestion. We present FOLD (Fuzzy Online Deduplication), an online fuzzy deduplication system that delivers high recall and throughput for evolving datasets. FOLD maintains an incrementally updated HNSW index over admitted documents, retrieving a small, high-quality candidate neighborhood for each incoming document instead of repeatedly rebuilding global buckets or rescanning the accumulated corpus. To our knowledge, FOLD is the first online fuzzy deduplication system to use HNSW. However, applying Jaccard similarity out of the box causes score crowding, making graph traversal unreliable within a small number of steps. FOLD addresses this with a bitmap representation that provides a more discriminative, Jaccard-aligned signal during HNSW search. Across four LLM-scale datasets (LM1B, C4, RealNews, and Common Crawl), FOLD stays fast and accurate as the corpus grows: at the largest evaluated scales, it maintains 93-97% recall and achieves up to 2.09x higher throughput than competing alternatives, whose best recall reaches only 76%.

David Wood, Boris Lublinsky, Alexy Roytman et al.

Data preparation is the first and a very important step towards any Large Language Model (LLM) development. This paper introduces an easy-to-use, extensible, and scale-flexible open-source data preparation toolkit called Data Prep Kit (DPK). DPK is architected and designed to enable users to scale their data preparation to their needs. With DPK they can prepare data on a local machine or effortlessly scale to run on a cluster with thousands of CPU Cores. DPK comes with a highly scalable, yet extensible set of modules that transform natural language and code data. If the user needs additional transforms, they can be easily developed using extensive DPK support for transform creation. These modules can be used independently or pipelined to perform a series of operations. In this paper, we describe DPK architecture and show its performance from a small scale to a very large number of CPUs. The modules from DPK have been used for the preparation of Granite Models [1] [2]. We believe DPK is a valuable contribution to the AI community to easily prepare data to enhance the performance of their LLM models or to fine-tune models with Retrieval-Augmented Generation (RAG).

Constantin Adam, Muhammed Fatih Bulut, Daby Sow et al.

Modern organizations struggle with insurmountable number of vulnerabilities that are discovered and reported by their network and application vulnerability scanners. Therefore, prioritization and focus become critical, to spend their limited time on the highest risk vulnerabilities. In doing this, it is important for these organizations not only to understand the technical descriptions of the vulnerabilities, but also to gain insights into attackers' perspectives. In this work, we use machine learning and natural language processing techniques, as well as several publicly available data sets to provide an explainable mapping of vulnerabilities to attack techniques and threat actors. This work provides new security intelligence, by predicting which attack techniques are most likely to be used to exploit a given vulnerability and which threat actors are most likely to conduct the exploitation. Lack of labeled data and different vocabularies make mapping vulnerabilities to attack techniques at scale a challenging problem that cannot be addressed easily using supervised or unsupervised (similarity search) learning techniques. To solve this problem, we first map the vulnerabilities to a standard set of common weaknesses, and then common weaknesses to the attack techniques. This approach yields a Mean Reciprocal Rank (MRR) of 0.95, an accuracy comparable with those reported for state-of-the-art systems. Our solution has been deployed to IBM Security X-Force Red Vulnerability Management Services, and in production since 2021. The solution helps security practitioners to assist customers to manage and prioritize their vulnerabilities, providing them with an explainable mapping of vulnerabilities to attack techniques and threat actors

Hajar Emami Gohari, Swanand Ravindra Kadhe, Syed Yousaf Shah et al.

Data quantity and quality play a vital role in determining the performance of Large Language Models (LLMs). High-quality data, in particular, can significantly boost the LLM's ability to generalize on a wide range of downstream tasks. Large pre-training datasets for leading LLMs remain inaccessible to the public, whereas many open datasets are small in size (less than 5 trillion tokens), limiting their suitability for training large models. In this paper, we introduce GneissWeb, a large dataset yielding around 10 trillion tokens that caters to the data quality and quantity requirements of training LLMs. Our GneissWeb recipe that produced the dataset consists of sharded exact sub-string deduplication and a judiciously constructed ensemble of quality filters. GneissWeb achieves a favorable trade-off between data quality and quantity, producing models that outperform models trained on state-of-the-art open large datasets (5+ trillion tokens). We show that models trained using GneissWeb dataset outperform those trained on FineWeb-V1.1.0 by 2.73 percentage points in terms of average score computed on a set of 11 commonly used benchmarks (both zero-shot and few-shot) for pre-training dataset evaluation. When the evaluation set is extended to 20 benchmarks (both zero-shot and few-shot), models trained using GneissWeb still achieve a 1.75 percentage points advantage over those trained on FineWeb-V1.1.0.