Trial by Trace
Benchmarking AI agent harnesses for reliability and failure attribution. TraceTrial-Bench examines 95 tasks across four harnesses and nine model configurations.
TRACELITE AI / RESEARCH TIMELINE
Technical reports, papers, and milestones. Follow our work in agent evaluation, synthetic data, and training as it develops.
Explore the timelineNEWEST FIRST
2026
-
Report dated
-
Article published
TraceLite AI in InfoQ’s AICon Shenzhen recap
InfoQ’s conference recap lists TraceLite AI among the organizations represented by speakers at AICon Shenzhen 2026.
-
Proceedings
SimRPD
Optimizing recruitment proactive dialogue agents through simulator-based data evaluation and selection.
-
Proceedings
Look Less, Reason More
Rollout-Guided Adaptive Pixel-Space Reasoning.
-
Conference year
MM-VeriRubric: Verifiable Evidence-Grounded Rubrics for Multimodal Reward Models
Grounding multimodal reward-model rubrics in task-relevant visual evidence and verifying their consistency.
-
Conference year
Disentangled Privileged States for Simulator-Assisted Multi-Turn Dialogue Policy Training
Separating privileged state tracking from simulated user utterances to support multi-turn dialogue policy training.
-
Conference year
Absorbing Gradient Conflicts: Modeling Semantic Variance via Kent Distributions for Cross-Modal Hashing
Modeling class proxies as Kent distributions on the hypersphere so that conflicting multi-label gradients are absorbed by directional variance instead of shifting the proxy.
-
Conference year
Select Less, Reason More: Prioritizing Evidence Purity for Video Reasoning
Selecting relevant frames and resampling visual evidence for video reasoning.
-
Conference year
CausalStep
A Benchmark for Explicit Stepwise Causal Reasoning in Videos.
No entries in this view.
Try another type or year, or explore the full timeline.