Search papers, labs, and topics across Lattice.
4
0
3
5
Converting noisy, human-centric guides into self-evolving agent skills can yield performance improvements of up to 25.3 percentage points across diverse tasks.
TVIR-Agent reveals that integrating visual elements into report generation can dramatically improve the quality and reliability of analytical outputs.
Current research agents still struggle with retrieval robustness and hallucination control, even when evaluated in a static, verifiable research environment.
Failure-driven post-training, combined with a meticulously curated 10M token STEM dataset, unlocks a 4.68% performance boost in LLM reasoning, proving that strategic data synthesis around model weaknesses is a powerful path to improvement.
