EPISODE · Jan 25, 2025 · 14 MIN
#3 – Kimi K1.5: Scaling Reinforcement Learning with LLMs
from Artificially Speaking · host Henry Moran
This technical report details the development and evaluation of Kimi k1.5, a multi-modal large language model (LLM) trained using reinforcement learning (RL). The researchers emphasize a novel approach focusing on long-context scaling and improved policy optimization, achieving state-of-the-art results on various benchmarks. Key innovations include a simplistic RL framework avoiding complex techniques, effective long2short methods to enhance short-CoT models, and infrastructure optimizations like partial rollouts for efficient training. The report thoroughly explores the RL training techniques, data curation strategies, and system architecture. Extensive experimental results across text, reasoning, and vision benchmarks demonstrate Kimi k1.5's superior performance.
What this episode covers
This technical report details the development and evaluation of Kimi k1.5, a multi-modal large language model (LLM) trained using reinforcement learning (RL). The researchers emphasize a novel approach focusing on long-context scaling and improved policy optimization, achieving state-of-the-art results on various benchmarks. Key innovations include a simplistic RL framework avoiding complex techniques, effective long2short methods to enhance short-CoT models, and infrastructure optimizations like partial rollouts for efficient training. The report thoroughly explores the RL training techniques, data curation strategies, and system architecture. Extensive experimental results across text, reasoning, and vision benchmarks demonstrate Kimi k1.5's superior performance.
NOW PLAYING
#3 – Kimi K1.5: Scaling Reinforcement Learning with LLMs
No transcript for this episode yet
Similar Episodes
Jun 1, 2025 ·89m
Apr 6, 2025 ·110m
Mar 31, 2025 ·11m
Mar 23, 2025 ·87m
Mar 16, 2025 ·55m
Mar 9, 2025 ·90m