The paper introduces RAG-Safety-Bench, a benchmark isolating how retrieval-augmented generation affects LLM safety by testing four controlled conditions: no-RAG baseline…
The paper introduces LOCUS, a post-training method that curbs verbosity in large language models by applying task-aware low-rank adaptations that…
Researchers present an enhanced quantization method for large language models that improves on Kashin-decomposition-based approaches by replacing dense random orthogonal…
The post describes a three-layer stack vLLM uses to optimize serving for agentic workloads: a data plane with hybrid KV…
vLLM’s tiered KV cache offloading system preserves evicted key-value cache data across host memory, storage, and remote peers instead of…
The paper details training Nemotron 3 Ultra with supervised fine-tuning and reinforcement learning to produce natural-language mathematical proofs, built around…
The paper studies how 13 different language model checkpoints degrade under recursive training on model-generated text, using a shared corpus…
This PyTorch blog post details how the Hugging Face Kernels project added support for Helion, a tile-based DSL from Meta…
TurboQuant is a vector-embedding compression technique that applies a random orthogonal rotation to embedding coordinates, then maps the rotated values…
A Hugging Face contributor describes a data-driven quantization scheme for GGUF models that empirically measures per-tensor sensitivity rather than relying…
The paper studies how large language models retrieve and use internal knowledge by performing layerwise interventions on hidden states across…
The paper proposes a Bayesian backward-reasoning method for aggregating decisions from multiple LLM agents, constructing reverse posteriors via explicit likelihood…
