3 citations · 3 across the 1 of their papers we have counts for
3 papers · 1 filter
LiveBench: A Challenging, Contamination-Limited LLM Benchmark
Colin White, Samuel Dooley, Manley Roberts +15
Test set contamination, wherein test data from a benchmark ends up in a newer model's training set, is a well-documented obstacle for fair LLM evaluation and can quickly render ben…
Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive
Arka Pal, Deep Karkhanis, Samuel Dooley +3
Direct Preference Optimisation (DPO) is effective at significantly improving the performance of large language models (LLMs) on downstream tasks such as reasoning, summarisation, a…
Data Contamination Through the Lens of Time
Manley Roberts, Himanshu Thakur, Christine Herlihy +2
Recent claims about the impressive abilities of large language models (LLMs) are often supported by evaluating publicly available benchmarks. Since LLMs train on wide swaths of the…