The Archive

Study evaluates LLM performance on automated qualitative coding of firearm violence survivor interviews, examining both capability gains and ethical risks.

Chain-of-thought prompting degrades spatial reasoning in multimodal LLMs; identifies shortcut learning and hallucination.

LLMSniffer detects LLM-generated code using GraphCodeBERT and contrastive learning on GPTSniffer/Whodunit datasets.

Mind's Eye benchmark evaluates multimodal LLMs on visual abstraction, relation, transformation via A-R-T taxonomy.

Survey of intrinsic interpretability architectures for LLMs; systematizes design principles for transparent model internals.

Discussion of whether manual programming remains viable as LLMs handle code generation; debates learning depth vs. on-demand intelligence.

Non-native English speaker reports mimicking LLM writing patterns after increased interaction with AI systems for learning.

Training LLMs requires periodic checkpoints. These full snapshots of model weights, optimizer states, and gradients are saved to storage so training can resume... Training LLMs requires periodic checkpoints. These full snapshots of model weights, optimizer states, and gradients…

Import AI 450 covers Chinese electronic warfare model, LLM psychological effects research, and cyberattack scaling laws.

Import AI 449 reports LLM-trained LLMs, 72B parameter distributed training, and vision vs. text scaling comparisons.

Berkeley BAIR develops scalable interaction identification methods to improve LLM interpretability and safety through feature/data attribution analysis.

Computer-aided engineering (CAE) is shifting from human-driven workflows toward AI-driven ones, including physics foundation models that generalize across... Computer-aided engineering (CAE) is shifting from human-driven workflows toward AI-driven ones, including physics…

Physical AI is rapidly evolving, from next-generation software-defined autonomous vehicles (AVs) to humanoid robots. The challenge is no longer how to run a... Physical AI is rapidly evolving, from next-generation software-defined autonomous vehicles (AVs) to humanoid robots…

OpenAI's IH-Challenge training method improves instruction hierarchy and prompt injection resistance in frontier LLMs.

Deploying large language models (LLMs) requires large-scale distributed inference, which spreads model computation and request handling across many GPUs and... Deploying large language models (LLMs) requires large-scale distributed inference, which spreads model computation and…

Deploying and optimizing large language models (LLMs) for high-performance, cost-effective serving can be an overwhelming engineering problem. The ideal... Deploying and optimizing large language models (LLMs) for high-performance, cost-effective serving can be an overwhelming…

Organizations deploying LLMs are challenged by inference workloads with different resource requirements. A small embedding model might use only a few gigabytes... Organizations deploying LLMs are challenged by inference workloads with different resource requirements. A small…

Import AI 446 covers nuclear power LLMs, China's comprehensive AI benchmark suite, and measurement for policy.

NVIDIA TensorRT LLM enables developers to build high-performance inference engines for large language models (LLMs), but deploying a new architecture... NVIDIA TensorRT LLM enables developers to build high-performance inference engines for large language models (LLMs), but…

Large language models (LLMs) are rapidly expanding their context windows, with recent models supporting sequences of 128K tokens, 256K tokens, and beyond.... Large language models (LLMs) are rapidly expanding their context windows, with recent models supporting sequences of…

In 2025, NVIDIA partnered with Black Forest Labs (BFL) to optimize the FLUX.1 text-to-image model series, unlocking FP4 image generation performance on NVIDIA... In 2025, NVIDIA partnered with Black Forest Labs (BFL) to optimize the FLUX.1 text-to-image model series, unlocking…

We keep seeing LLMs with larger context windows in the news, along with promises that they can hold entire conversation histories, volumes of books, or multiple... We keep seeing LLMs with larger context windows in the news, along with promises that they can hold entire…

Large language models (LLMs) and multimodal reasoning systems are rapidly expanding beyond the data center. Automotive and robotics developers increasingly want... Large language models (LLMs) and multimodal reasoning systems are rapidly expanding beyond the data center…

OpenAI introduces real-world testing methods to measure and reduce political bias in ChatGPT.

Writing code is only one part of effective software engineering.