Manav Singhal

BrowserGym: Technical deep dive into web agent automation

BrowserGym: Technical deep dive into web agent automation

The field of web automation faces significant challenges in standardizing agent development and evaluation. BrowserGym, a Gym environment for web automation tasks by ServiceNow, addresses these challenges by providing a unified framework that standardizes the development, testing, and evaluation of web agents. In addition, they also design AgentLab, a complementary

Image generated using Meta AI

Ensuring responsible AI: An overview of DeepMind’s FACTS framework

Introduction DeepMind introduces the FACTS Grounding, an online leaderboard and benchmark designed to evaluate the factual accuracy of language models' long-form responses based on a given context in user prompts. The benchmark requires models to generate text grounded in a provided document, which can be up to 32,000

Image credit: Meta

Innovative training of LLMs in continuous latent spaces, by Meta AI

LLMs have made significant progress in tasks based on language processing and understanding. However, their reasoning capabilities, particularly in complex scenarios, often fall short. Traditional approaches like Chain-of-Thought (CoT) reasoning, which asks the model to reason before answering by thinking step-by-step, face inherent limitations. A recent paper introduces a novel

Inside OpenAI’s o1: Part 2

Inside OpenAI’s o1: Part 2

As we have discussed in our last blog, the o1 family is trained to think before it gives an output to mimic the way humans think for some time before giving an answer to aid reasoning. Beyond the traditional evaluations, we shall go through a few vibe check evals done

Inside OpenAI’s o1: Part 1

Inside OpenAI’s o1: Part 1

Introduction The o1 family of models is trained using reinforcement learning to perform complex reasoning. It is baking in the now familiar chain of thoughts before giving the output. Through the training, the models learn to refine their thinking, explore different strategies, and identify mistakes in those exploration paths. It

Data contamination

Evaluating data contamination in LLMs

Introduction In recent years, large language models (LLMs) like GPT and Llama have constantly been pushing accuracy numbers on popular benchmarks. However, one issue that has become increasingly important is data contamination—the overlap between a model’s training data and the evaluation benchmarks used to assess its performance. This

System overview

Red teaming with auto-generated rewards and multi-step RL

Introduction Making AI systems like LLMs robust against adversarial cases is a critical area of research. One approach to identifying vulnerabilities in AI models is red-teaming, where adversarial prompts or attacks are designed to expose weaknesses. However, generating a wide variety of diverse yet effective attacks remains a significant challenge.