Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories
Paper • 2606.02060 • Published • 40 •
3
Avi
avahal
AI & ML interests
LLMs
Recent Activity
commentedon a paper about 3 hours ago
Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories commentedon a paper about 5 hours ago
Trust Region On-Policy Distillation commentedon a paper about 5 hours ago
From Activation to Causality: Discovery of Causal Visual Representations in the Human BrainOrganizations
None yet
commented a
paper about 3 hours ago
commented 3
papers about 5 hours ago
Trust Region On-Policy Distillation
Paper • 2606.01249 • Published • 37 •
2
From Activation to Causality: Discovery of Causal Visual Representations in the Human Brain
Paper • 2605.23895 • Published • 50 •
2
OCC-RAG: Optimal Cognitive Core for Faithful Question Answering
Paper • 2606.00683 • Published • 81 •
6
commented 8
papers about 15 hours ago
A Matter of TASTE: Improving Coverage and Difficulty of Agent Benchmarks
Paper • 2605.28556 • Published • 61 •
3
Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decoding
Paper • 2605.29707 • Published • 135 •
3
On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters
Paper • 2606.02437 • Published • 168 •
4
Crafter: A Multi-Agent Harness for Editable Scientific Figure Generation from Diverse Inputs
Paper • 2605.30611 • Published • 184 •
3
SwanVoice: Expressive Long-Form Zero-Shot Speech Synthesis for Both Monologue and Dialogue
Paper • 2605.30993 • Published • 56 •
3
Trust-Region Behavior Blending for On-Policy Distillation
Paper • 2605.31159 • Published • 64 •
4
GrepSeek: Training Search Agents for Direct Corpus Interaction
Paper • 2605.29307 • Published • 98 •
5
COLLEAGUE.SKILL: Automated AI Skill Generation via Expert Knowledge Distillation
Paper • 2605.31264 • Published • 104 •
3
commented 4
papers 8 days ago
MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research
Paper • 2605.26114 • Published • 64 •
3
SpatialBench: Is Your Spatial Foundation Model an All-Round Player?
Paper • 2605.27367 • Published • 71 •
4
EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video Generation
Paper • 2605.23271 • Published • 79 •
3
LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding
Paper • 2605.27365 • Published • 137 •
4
commented 4
papers 9 days ago
Seeing the Needle in the Haystack: Towards Weakly-Supervised Log Instance Anomaly Localization via Counterfactual Perturbation
Paper • 2605.10988 • Published • 3 •
4
Decoding the Critique Mechanism in Large Reasoning Models
Paper • 2603.16331 • Published •
3
ClaimDiff-RL: Fine-Grained Caption Reinforcement Learning through Visual Claim Comparison
Paper • 2605.20278 • Published • 1 •
3
Pixel-Level Pavement Distress Assessment Using Instance Segmentation
Paper • 2605.26095 • Published •
2