Trace
PulseAugur coverage of Trace — every cluster mentioning Trace across labs, papers, and developer communities, ranked by signal.
- 2026-06-10 research_milestone Researchers introduced the TRACE method for detecting LLM ghostwriters, achieving state-of-the-art performance on a new dataset. source
- 2026-06-02 research_milestone A new framework called TRACE was introduced, significantly improving multi-video event understanding and claim generation. source
11 day(s) with sentiment data
-
New TRACE framework detects RAG poisoning attacks via token influence
Researchers have developed a new framework called TRACE to detect poisoning attacks in retrieval-augmented generation (RAG) systems. These attacks manipulate RAG models by inserting malicious documents into their retrie…
-
New AI Research Unveils Methods for Understanding Neural Network Computation
Two new research papers introduce novel methods for understanding the internal workings of complex neural networks. The first, TRACE, proposes a new paradigm for learning to compute on circuit graphs by using a Hierarch…
-
Software Studio Fervon Launches, Leveraging AI Agents for Development
Jon Martin has launched Fervon, a new software studio that will build products using AI agents. The studio's portfolio includes Trace and various open-source development tools, all developed with the assistance of AI agents.
-
Trace launches as an on-device, privacy-focused meeting transcription tool
Trace is a new macOS application designed for on-device meeting transcription and note-taking. It operates entirely locally, ensuring that audio and transcripts never leave the user's Mac, addressing privacy concerns as…
-
New research tackles spoofed speech detection with advanced AI models
Researchers are developing advanced methods to detect spoofed speech, a growing challenge due to realistic synthesis and voice conversion technologies. One approach, the Temporal Pyramid Adapter, uses parallel temporal …
-
New TRACE memory framework aids robots in delayed-evidence visuomotor tasks
Researchers have developed TRACE (TRAjectory-routed Causal Evidence), a novel memory framework designed to help robots make decisions based on past visual information that is no longer visible. This system uses path sig…
-
New TRACE method detects LLM ghostwriters in long texts
Researchers have developed a new method called TRACE to detect ghostwriters generated by large language models in long-form texts. This technique creates a unique fingerprint by analyzing token-level transition patterns…
-
New methods enhance AI agent reliability and safety
Researchers have developed new methods to improve the reliability and safety of AI agents. One approach, TRACE, focuses on monitoring long-horizon agent trajectories to detect malicious or unintended behaviors by analyz…
-
TRACE framework uses autoregressive models for causal discovery
Researchers have developed a new framework called TRACE that leverages autoregressive models to uncover causal relationships within sequential data. This method repurposes existing language models to perform causal disc…
-
TRACE framework boosts multi-video event understanding with evidence grounding
Researchers have developed TRACE, a new framework designed to improve multi-video event understanding and claim generation. TRACE employs a ground-before-reasoning strategy, first creating text-searchable timelines for …
-
New TRACE method improves safety detection for long-horizon LLM agents
Researchers have introduced TRACE, a novel method for enhancing the safety of long-horizon Large Language Model (LLM) agents. TRACE addresses the challenge of detecting sparse and delayed safety risks that are often mis…
-
New research tackles machine unlearning challenges in LLMs
Researchers are developing new methods for machine unlearning in large language models, a process crucial for privacy and knowledge management. Several papers explore techniques to remove specific data from trained mode…
-
TRACE method discovers task-specific LLM parameters to prevent forgetting
Researchers have developed a new method called TRACE for continual fine-tuning of large language models. This approach identifies and isolates a small subset of task-specific parameters, updating only those while keepin…
-
New On-Policy Replay method combats LLM forgetting
Researchers have developed a new method called On-Policy Replay (OPR) to address catastrophic forgetting in large language models during continual supervised fine-tuning. OPR filters historical prompts based on a task r…
-
AI Agent TRACE Enhances Drug Discovery Lead Optimization
Researchers have developed TRACE, a novel agent that utilizes LLM reasoning for molecular lead optimization in drug discovery. Unlike previous methods that optimize in a single step, TRACE treats tool selection as a seq…
-
New AI methods tackle complex inverse problems with improved sampling
Researchers are developing new methods to tackle complex inverse problems in machine learning, particularly in scenarios where gradient information is unavailable. New techniques aim to improve sampling from high-dimens…
-
New synthetic dataset TRACE aids AI training in Applied Behavior Analysis
Researchers have developed TRACE, a synthetic dataset designed to train AI models in Applied Behavior Analysis (ABA). This dataset comprises 2,999 examples covering teaching-program generation and multi-session behavior…
-
New TRACE benchmark evaluates trustworthy tourism AI recommendations
Researchers have introduced TRACE, a new benchmark dataset designed to evaluate conversational recommender systems in the tourism domain. TRACE addresses the need for systems that not only suggest relevant points of int…
-
TRACE framework enhances conformal prediction with diffusion and flow matching
Researchers have introduced TRACE, a novel framework for conformal prediction designed to handle multi-dimensional outputs. This method defines nonconformity by aligning transport dynamics within diffusion and flow matc…
-
TRACE model predicts delayed conversion rates using post-click behavior
Researchers have introduced TRACE, a novel method for predicting online conversion rates when feedback is delayed. TRACE addresses the challenge by analyzing post-click user behavior trajectories rather than relying sol…