DeepSeek-R-1-Distill-14B

'DeepSeek-R-1-Distill-14B' is a benchmark dataset used to evaluate the efficiency of language reasoning in large language models by assessing their performance on diverse tasks while minimizing output token count.

Papers using DeepSeek-R-1-Distill-14B (2)

TL;DR: Too Long, Do Re-weighting for Efficient LLM Reasoning Compression2025TL;DR: Too Long, Do Re-weighting for Effcient LLM Reasoning Compression2025