Large Language Models

New benchmark for evaluating long-context models finds sub-O(n) methods lacking in real-world use cases.

BLT: tokenizer-free LLM for efficiency and robustness

Norwegians show that using copyrighted material improves LLMs, but raises legal and ethical issues.

RULEARENA, a new benchmark, rigorously evaluates large language models’ ability to apply complex, real-world rules across diverse scenarios, revealing significant shortcomings in current LLMs’ rule-gu…

Phi-4: a 14B parameter LLM surpassing its teacher model (GPT-4) in STEM-focused QA through innovative synthetic data generation and post-training techniques.

JuStRank: LLM system ranker benchmark reveals critical judge qualities (decisiveness, bias) impacting ranking accuracy, highlighting instance-level performance doesn’t guarantee accurate system-level…

Fine-tuning small language models? Tweak the learning rate and batch size for a reasoning boost!

Granite Guardian: Open-source risk detection models for LLMs, surpassing existing models in accuracy and offering comprehensive coverage across multiple risk dimensions, promoting safer AI.

Contextualized AI counterspeech significantly outperforms generic methods by adapting to the moderation context and user, improving persuasiveness without sacrificing other qualities.

LLMs are trained to reason using language, but COCONUT lets them reason directly in a continuous latent space, boosting performance on logical tasks requiring complex planning.

Moxin-LLM: A fully open-source 7B parameter LLM achieving superior zero-shot performance, promoting transparency and reproducibility in AI research.

LG AI Research unveils EXAONE 3.5, a series of instruction-tuned language models (2.4B, 7.8B, and 32B parameters) excelling in real-world tasks, long-context understanding, and general benchmarks.

CodeArena, a novel benchmark, evaluates code LLMs based on human preferences, revealing performance gaps between open-source and proprietary models, and a large-scale synthetic instruction corpus impr…

MONET improves Transformer interpretability by using Mixture-of-Experts (MoE) with 262K monosemantic experts per layer, achieving parameter efficiency and enabling knowledge manipulation without perfo…

Marco-LLM: A groundbreaking multilingual LLM significantly enhances cross-lingual capabilities via massive multilingual training, bridging the performance gap between high- and low-resource languages.

LLMs’ training quality is exponentially improving, enabling models with half the parameters to match state-of-the-art performance every 3 months, thus reducing inference costs.

WRPO: Implicitly fuse LLMs, boosting performance without complex alignment or merging!

Researchers developed a robust multi-bit text watermarking method using LLMs for paraphrasing, achieving over 99.99% detection accuracy while maintaining semantic information and resisting common atta…

AGORABENCH: A new benchmark reveals surprising strengths & weaknesses of LMs as synthetic data generators, showing that problem-solving ability isn’t the sole indicator of data quality.

Imperfect OCR hinders Retrieval-Augmented Generation (RAG). OHRBench, a new benchmark, reveals this cascading impact, showing current OCR solutions insufficient for high-quality RAG knowledge bases. …