Search papers, labs, and topics across Lattice.
IISER Kolkata, India shubhra_2303res116@iitp.ac.in, {abhilekh.229301149, aditya.23fe10cds00348}@muj.manipal.edu, kripaghosh@iiserkol.ac.in
2
0
6
5
LLMs, despite their prowess, stumble and hallucinate when questions are subtly obfuscated, revealing a surprising fragility in their factual QA abilities.
Achieve up to 39.6% FLOP reduction in LLM inference without retraining or architectural changes using QuickSilver's dynamic token-level optimizations.
