Artificially Speaking

The provided text introduces  Cognitive Architectures for Language Agents (CoALA) , a theoretical framework designed to standardize the development of artificial intelligence systems that use large language models as their core. By drawing parallels between modern AI and historical  symbolic cognitive science , the authors propose a modular structure involving  working and long-term memory , a div...

SkillOpt  is a novel optimization framework designed to improve the performance of  AI agents  by treating their skills as a trainable, external text document. Unlike traditional methods that rely on manual prompting or model fine-tuning, this system uses a  separate optimizer model  to analyze successes and failures from task trajectories and propose structured edits. These edits are governed by...

The document introduces  Vending-Bench , a novel  simulated environment  designed to evaluate the  long-term coherence of autonomous agents  powered by Large Language Models (LLMs). The benchmark tasks agents with managing a  virtual vending machine business  over extended periods, requiring them to handle  inventory, pricing, ordering, and daily fees . While some advanced LLMs like Claude 3.5 Son...

This research paper explores the capabilities of large language models (LLMs) in competitive programming. It compares the performance of OpenAI's o1 and o3 LLMs, highlighting the significant improvement in performance achieved by o3 through increased reinforcement learning. The study also examines a specialized LLM, o1-ioi, fine-tuned for the International Olympiad in Informatics (IOI), demons...

This research paper explores using executable Python code as actions for Large Language Model (LLM) agents. The authors introduce CodeAct, a framework enabling LLMs to generate and execute Python code, dynamically adapting actions based on observations. Experiments across 17 LLMs demonstrate CodeAct's superior performance in complex tasks, achieving up to a 20% higher success rate than alterna...

This research paper investigates the effectiveness of ensembling different large language models (LLMs) to improve performance. The authors introduce Self-MoA, a method that aggregates multiple outputs from a single, top-performing LLM, contrasting it with traditional Mixture-of-Agents (MoA) which combines outputs from multiple LLMs. Experiments across various benchmarks show Self-MoA significantl...

This technical report introduces the Qwen2.5-1M series of large language models, significantly enhancing long-context capabilities (up to 1 million tokens) through novel training and inference techniques. Key improvements involve long data synthesis, progressive pre-training, and a multi-stage fine-tuning process. The report also details a newly open-sourced inference framework featuring a length...

The Paper details the development and capabilities of MiniMax-01, a series of large language models (LLMs) and vision-language models (VLMs).  Key features  include superior long-context processing (up to 4 million tokens), utilization of lightning attention, and a Mixture of Experts (MoE) architecture.  Performance  is shown to be comparable to state-of-the-art models like GPT-4 and Claude. The p...

This research paper explores the development and evaluation of Process Reward Models (PRMs) for improving mathematical reasoning in Large Language Models (LLMs). The authors identify limitations in using Monte Carlo estimation for data annotation and Best-of-N evaluation for assessing PRM performance, proposing a consensus filtering mechanism combining Monte Carlo estimation with an LLM-as-a-judge...

This research paper explores using generative AI, specifically ChatGPT, to automate Test-Driven Development (TDD). The authors propose two interaction patterns: a collaborative approach where developers and AI work together, and a fully automated approach. An experiment comparing these patterns with traditional TDD revealed that while generative AI can improve efficiency, human supervision is cruc...

The paper introduces rStar-Math, a novel method that significantly improves the mathematical reasoning capabilities of small language models (SLMs). It achieves this through a self-evolving process using Monte Carlo Tree Search (MCTS) and three key innovations: code-augmented data synthesis, a novel process preference model (PPM) training method, and a four-round self-evolution recipe. rStar-Math...

Humanity's Last Exam (HLE)  is a new, extremely difficult, multi-modal benchmark designed to evaluate large language models (LLMs). Created by a global team of experts, HLE features 3,000 questions spanning numerous academic subjects, surpassing the capabilities of current LLMs. The benchmark's questions were rigorously tested and reviewed to ensure difficulty and eliminate easy solutions,...

Researchers from Mondragon University and the University of Seville conducted a pre-deployment safety evaluation of OpenAI’s o3-mini large language model (LLM). They used their tool, ASTRAL, to automatically generate 10,080 unsafe prompts, covering 14 safety categories. The study found 87 instances of unsafe LLM behavior after manual verification, highlighting the o3-mini's relatively high saf...

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 techni...

A deep dive into the DeepSeek-R1: Reasoning via Reinforcement Learning research paper. DeepSeek-R1, a large language model enhanced for reasoning capabilities through reinforcement learning (RL). Two versions are described: DeepSeek-R1-Zero, trained solely with RL, and DeepSeek-R1, which incorporates a multi-stage training process including cold-start data and supervised fine-tuning to improve rea...

This research paper introduces Titans, a novel family of neural architectures designed to improve long-term memory in sequence modeling. Titans incorporate a deep neural long-term memory module that learns to memorize historical context at test time, augmenting a short-term attention mechanism.