LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression (2024 · 161 citations)
Long prompts built with techniques like RAG, in-context learning or chain-of-thought increase the cost and latency of LLM inference.
MathFusion: Enhancing Mathematical Problem-solving of LLM through Instruction Fusion (2025 · 10 citations)
Most data augmentation for mathematical SFT changes one problem at a time, for example by rewording it or adjusting its difficulty.
ChartVerse: Scaling Chart Reasoning via Reliable Programmatic Synthesis from Scratch (2026 · 5 citations)
Open-source vision-language models lag on chart reasoning, mainly because training data is weak. Synthetic charts tend to be simple and repetitive, and automatically generated QA pairs often contain hallucinated or numerically wrong answers. ChartVerse is a code-driven pipeline that builds both charts and reasoning…
REST: Stress Testing Large Reasoning Models by Asking Multiple Problems at Once (2026 · 3 citations)
Standard reasoning benchmarks pose one question per prompt. Many, such as GSM8K and MATH500, are now close to saturated for large reasoning models (LRMs), which forces repeated, costly creation of new test sets. The single-question format also ignores real deployments where a model must handle several queries in one…

Co-authors here

12 of 22 co-authors with a page, the most papers together first.

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

Papers of our library cite their works 126 times (of the 149,439 papers whose bibliographies we read); 29 citing papers with a page here →.

Updated 1 November 2026