Ok, you have a moderately complex math problem you needed to solve. You gave the problem to 6 LLMS all paid versions. All 6 get the same numbers. Would you trust the answer?
Ok, you have a moderately complex math problem you needed to solve. You gave the problem to 6 LLMS all paid versions. All 6 get the same numbers. Would you trust the answer?
Probably, depending on the context. It is possible that all 6 models were trained on the same misleading data, but not very likely in general.
Number crunching isn’t an obvious LLM use case, though. Depending on the task, having it create code to crunch the numbers, or a step-by-step tutorial on how to derive the formula, would be my preference.