2 ms·
> On individual tasks Claude and GPT are comparable That is not what the first graphs show - the Anthropic models cluster at 'better' positions on the graph, a
by bonsai_spool 6mo ago
> On individual tasks Claude and GPT are comparable
That is not what the first graphs show - the Anthropic models cluster at 'better' positions on the graph, and I imagine you could show that the values are significantly different.