Can we obtain significant success in RST discourse parsing by using Large Language Models?

Benchmark Model Rank Results
discourse-parsing-on-rst-dtBottom-up Llama 2 (70B)#1Standard Parseval (Full): 58.1Standard Parseval (Span): 79.8…
discourse-parsing-on-rst-dtTop-down Llama 2 (70B)#2Standard Parseval (Full): 56.0Standard Parseval (Span): 78.8…
discourse-parsing-on-rst-dtBottom-up Llama 2 (13B)#3Standard Parseval (Full): 56.0Standard Parseval (Span): 78.3…
discourse-parsing-on-rst-dtBottom-up Llama 2 (7B)#4Standard Parseval (Full): 55.8Standard Parseval (Span): 78.2…
discourse-parsing-on-rst-dtTop-down Llama 2 (13B)#6Standard Parseval (Full): 55.6Standard Parseval (Span): 78.6…
discourse-parsing-on-rst-dtTop-down Llama 2 (7B)#13Standard Parseval (Full): 53.4Standard Parseval (Span): 76.3…