2026-06-24

Internal Generation Record

Internal generation metadata: 457 candidate papers.

Published 2026-06-24 Target source 2026-06-22 Actual source 2026-06-22 Candidates 457 Featured 6 Tracked 20

Generation Record

This page preserves selected papers, candidate scale, and source-date metadata for traceability. The page only changes presentation, not selected papers, ordering, or counts.

Internal generation record. Fetched at 2026-06-23T22:20:52.227535+00:00. Generated at 2026-06-23T22:22:20.400300+00:00. Machine-readable details stay under data/processed and data/reports.

Selected papers

RankTakeawayTopicarXiv
1strengthen multimodal understanding of charts, documents, and visual evidenceMultimodal Models2606.23063
2make agents use tools and reusable skills more reliablyBenchmarks and Evaluation2606.23032
3make agents use tools and reusable skills more reliablyBenchmarks and Evaluation2606.23678
4make agents use tools and reusable skills more reliablyAgents and Tool Use2606.23525
5improve model reasoning, planning, and verificationReasoning and Planning2606.23672
7make RAG retrieval and knowledge-base QA more reliableTraining and Post-training2606.23200
6improve model reasoning, planning, and verificationBenchmarks and Evaluation2606.23459
8improve code generation, execution feedback, and automated repairSpeech and Audio2606.23060
9make RAG retrieval and knowledge-base QA more reliableVideo Generation2606.23023
10make agents use tools and reusable skills more reliablyBenchmarks and Evaluation2606.22958
11improve model reasoning, planning, and verificationBenchmarks and Evaluation2606.22873
12improve code generation, execution feedback, and automated repairVideo Generation2606.22726
13improve code generation, execution feedback, and automated repairCode Intelligence2606.23545
14improve model reasoning, planning, and verificationRetrieval and RAG2606.23108
15make agents use tools and reusable skills more reliablyVision and Image Generation2606.22966
16improve code generation, execution feedback, and automated repairCode Intelligence2606.22946
17make RAG retrieval and knowledge-base QA more reliableCode Intelligence2606.22807
18make RAG retrieval and knowledge-base QA more reliableBenchmarks and Evaluation2606.22778
19strengthen multimodal understanding of charts, documents, and visual evidenceCode Intelligence2606.23668
20make agents use tools and reusable skills more reliablyAgents and Tool Use2606.23654
21make RAG retrieval and knowledge-base QA more reliableMultimodal Models2606.23643
22make RAG retrieval and knowledge-base QA more reliableRobotics and Embodied AI2606.23634
23make RAG retrieval and knowledge-base QA more reliableBenchmarks and Evaluation2606.23603
24make RAG retrieval and knowledge-base QA more reliableMultimodal Models2606.23593
25identify and reduce safety, jailbreak, and alignment risksBenchmarks and Evaluation2606.23583
26improve code generation, execution feedback, and automated repairTraining and Post-training2606.23566