OptScale introduces a probabilistic framework that determines the optimal number of inference samples needed to meet a target accuracy with a confidence guarantee, dramatically reducing token usage while maintaining or improving performance across various large‑language‑model benchmarks.
This article analyzes the RETuning framework, which addresses LLMs' bias toward analyst opinions and lack of evidence weighting in stock movement prediction by introducing a two‑stage cold‑start fine‑tuning and reinforcement learning pipeline, evaluating it on the large Fin‑2024 dataset and demonstrating significant F1 gains, inference‑time scaling, and out‑of‑distribution robustness.
SANA 1.5 introduces an efficient model‑growth pipeline, depth‑pruning, and inference‑time scaling that reuse a 1.6 B‑parameter foundation to train a 4.8 B model with 8× lower memory, 60 % less training time, and GenEval scores that rival or surpass much larger diffusion models.
This article examines how to build and improve reasoning‑capable large language models, explains the definition and use‑cases of reasoning models, details DeepSeek‑R1’s training pipeline, compares four key enhancement methods—including inference‑time scaling, pure RL, SFT + RL, and distillation—and offers budget‑friendly advice.
This article explains what reasoning language models are, outlines their strengths and weaknesses, details DeepSeek R1's three variants and their training pipelines—including pure reinforcement learning, SFT + RL, and distillation—while also discussing inference‑time scaling techniques and related research such as Sky‑T1 and TinyZero.
This article explains what reasoning language models are, when they are needed, and reviews four main techniques— inference‑time scaling, pure reinforcement learning, combined SFT + RL, and distillation—illustrated with DeepSeek‑R1’s development, cost analysis, and low‑budget alternatives.
This article examines DeepSeek R1’s three breakthroughs, its low‑cost optimizations that bypass CUDA, and the resulting impact on the AI ecosystem, then provides a detailed technical review of seven open‑source reproductions—Open‑R1, Tiny‑Zero, SimpleScaling‑S1, and simpleRL‑reason—covering their architectures, reinforcement‑learning pipelines, and code implementations.
This report details the design, data synthesis, three‑stage training pipeline, and benchmark evaluation of the open‑source Steiner reasoning model, which aims to emulate OpenAI o1’s inference‑time scaling while highlighting current performance gaps and future research challenges.
