A running collection of published research, briefings, and analysis from this contributor.
StepFly achieves 94% success rate automating troubleshooting guides with AI agents, reducing execution time by 32.9-70.4% through parallel processing and DAG workflows.
Read the briefingComprehensive evaluation of GPT-4, GPT-4o, Gemini 1.5 Pro, DeepSeek-V3, and other LLMs across three core social media analytics tasks on...
Qwen3.5-Omni scales to hundreds of billions of parameters with 256k context length, achieving SOTA results across 215 audio-visual benchmarks and...
IndiaFinBench evaluates large language models on Indian financial regulatory text with 406 expert-annotated questions from SEBI and RBI documents.
LegalBench-BR introduces the first public benchmark for evaluating language models on Brazilian legal text classification with 3,105 appellate proceedings from...
LLHKG framework enables lightweight language models to construct knowledge graphs with performance comparable to GPT-3.5, automating entity and relation extraction.
StepFly automates troubleshooting guides for IT incidents using AI agents, achieving 94% success rate with 32.9-70.4% faster execution through parallel...
Researchers evaluated GPT-4, Gemini 1.5 Pro, and other LLMs across three social media analytics tasks using Twitter data, establishing new...
IndiaFinBench introduces 406 expert-annotated question-answer pairs from SEBI and RBI documents to evaluate large language model performance on Indian financial...
LegalBench-BR introduces the first public benchmark for evaluating large language models on Brazilian legal text classification with 3,105 court proceedings.
