squad
PulseAugur coverage of squad — every cluster mentioning squad across labs, papers, and developer communities, ranked by signal.
- 2026-10-28 product_launch The Squad AI agent team framework released version 2.0.0-preview.2, introducing several improvements and new features. source
8 day(s) with sentiment data
-
RAG pipeline safety fails under adversarial testing
A recent evaluation of a retrieval-augmented generation (RAG) pipeline revealed significant vulnerabilities in its safety mechanisms. The system, designed with hybrid retrieval and a relevance threshold, failed to preve…
-
Squad agent framework releases preview version 2.0.0-preview.3
The Squad project has released version 2.0.0-preview.3, an update to its agent framework. This release is available on GitHub and is associated with GitHub Copilot, indicating its use in software development contexts. T…
-
Squad AI agent framework releases v2.0.0-preview.2 with key improvements
The Squad AI agent team framework has released version 2.0.0-preview.2. This update introduces several improvements, including better npm reconciliation, enhanced parsing for skill frontmatter, and state collision preve…
-
AI models fail to detect fabricated facts in questions, research finds
A new research paper highlights a critical flaw in how current AI models handle unanswerable questions. These models can be tricked into using fabricated information presented within a question, even if the source mater…
-
New research reveals LLMs struggle to verify information, accepting 'usable falsehoods'
A new paper explores how large language models struggle with verifying information, particularly when unsupported premises are presented in a usable format. Researchers found that models are more likely to provide an an…
-
LLM evaluation methods significantly skew results, impacting fairness and RAG.
Recent developments in LLM evaluation highlight how subtle changes in methodology can significantly impact results. Adjusting rating scales, even by simple numerical changes, can alter LLM judge verdicts by approximatel…
-
New RAG benchmarking framework optimizes LLM configuration
Researchers have developed RAG-NIMBF, a new benchmarking framework designed to help optimize the configuration of retrieval-augmented generation (RAG) systems. This modular framework attaches to existing RAG pipelines, …
-
New research links model uncertainty to RAG hallucination risk
A new arXiv paper explores the relationship between uncertainty and hallucination risk in retrieval-augmented generation (RAG) models. The research indicates that uncertainty can effectively measure hallucination risk, …
-
ScaleDown launches Constrained Decoding Models for verifiable AI outputs
ScaleDown has introduced four new Constrained Decoding Models (CDMs) designed to improve the reliability and efficiency of AI tasks. Unlike traditional LLMs that generate answers token by token, CDMs are trained for spe…
-
LLM Caching: Faster local starts, higher Claude bills
A developer explored the impact of various caching strategies on LLM performance and cost. On a local Apple MacBook Air running a 4B model via Ollama, retrieval and prompt caching reduced the time to first token by appr…
-
AI Agents: Squad v1.0.1 Released, A2A Protocol Launches CLI
Squad v1.0.1 has been released, offering new capabilities for AI agents and software development, with integration for GitHub Copilot. Separately, the A2A Protocol has introduced the A2A CLI, a command-line tool designe…
-
Laya-compactor cuts RAG context tokens by 70% locally
Laya-compactor is a new open-source tool designed to reduce the number of tokens sent to large language models in Retrieval-Augmented Generation (RAG) pipelines. It works by scoring and filtering retrieved documents loc…
-
User trained small LLM with custom dataset 1-2 years ago
A user reflects on training a small LLM approximately one to two years ago using a custom dataset comprising their Mastodon posts, personal writings, code, and Q&A factoids, supplemented by open datasets like SQuAD. The…
-
SQuad framework slashes Video Transformer compute costs with sub-quadratic attention
Researchers have developed SQuad, a Sub-Quadratic Attention Distillation framework designed to improve the efficiency of Video Diffusion Transformers (DiTs). This new method reduces the computational cost of the self-at…
-
BERT-based QA models assessed for reliability; RoBERTa shows most stability
A new study published on arXiv evaluates the reliability of several BERT-based models, including RoBERTa, ALBERT, and DistilBERT, when applied to question-answering tasks. Researchers assessed model stability by introdu…
-
New RAG verification method improves multi-hop question answering
A new research paper proposes a novel approach to improve verification in retrieval-augmented generation (RAG) systems, particularly for multi-hop question answering. The study demonstrates that traditional per-chunk fi…
-
New cMoLLM architecture scales LLMs via dynamic convolutions
Researchers have introduced cMoLLM, a novel approach to scaling large language models by incorporating a mixture-of-experts (MoE) style throughout the entire model pipeline, rather than just in the feed-forward networks…
-
Microsoft details Agent Framework for disposable AI agents
Microsoft has detailed its Agent Framework, a layered SDK designed for building AI agents. The framework emphasizes disposable agents that can leverage durable memory, enabling complex task execution. This architecture …
-
New RAG research separates context length from semantic competition
A new research paper proposes a method to distinguish between context length and semantic competition as causes for errors in retrieval-augmented generation (RAG) systems. The study introduces a matched-control protocol…
-
New AS-LoRA method improves privacy in federated learning
Researchers have developed AS-LoRA, a novel framework for adaptive selection of LoRA components in privacy-preserving federated learning. This method addresses aggregation errors common in such setups by allowing each l…