Language Model
49 articles in this category (Page 2 of 3)
A Coding Implementation to Automating LLM Quality Assurance with DeepEval, Custom Retrievers, and LLM-as-a-Judge Metrics
This article details a coding implementation for automated LLM quality assurance, achieving rigorous testing through DeepEval, custom retrievers, and LLM-as-a-judge metrics.
TII Abu-Dhabi Released Falcon H1R-7B: A New Reasoning Model Outperforming Others in Math and Coding
Technology Innovation Institute (TII) released Falcon-H1R-7B, a 7B parameter model achieving performance comparable to 14B-47B models in math, code, and reasoning benchmarks.
Zhipu AI Releases GLM-4.6V: A 128K Context Vision Language Model with Native Tool Calling
Zhipu AI launched GLM-4.6V, a 106B parameter multimodal model with a 128K token context window, enabling native multimodal function calling for improved agent capabilities.
Apple Researchers Release CLaRa: A Continuous Latent Reasoning Framework for Compression-Native RAG with 16x–128x Semantic Document Compression
Apple's CLaRa achieves 16x–128x semantic document compression, boosting RAG efficiency without sacrificing accuracy.
Google AI Introduces Consistency Training for Safer Language Models Under Sycophantic and Jailbreak Style Prompts
Google AI introduces Consistency Training (Bias Augmented Consistency Training and Activation Consistency Training) to enhance language models' safety against sycophantic and jailbreak prompts while preserving their capabilities.
Anthropic's Research Demonstrates Claude's Introspective Awareness Through Concept Injection in Controlled Layers
Anthropic's study reveals that Claude models can detect injected concepts via internal activations, offering causal evidence of introspection. The research highlights controlled success rates and implications for LLM transparency.
Google AI Unveils Supervised Reinforcement Learning (SRL): A Step-Wise Framework for Enhancing Small Language Models
Google AI introduces Supervised Reinforcement Learning (SRL), a novel training framework that improves small language models' reasoning capabilities by leveraging expert trajectories and step-wise reward mechanisms.