From the file: "Answer is always line 1. Reasoning comes after, never before."
LLMs are autoregressive (filling in the completion of what came before), so you'd better have thinking mode on or the "reasoning" is pure confirmation bias seeded by the answer that gets locked in via the first output tokens.
Yeah this seems to be a very bad idea. Seems like the author had the right idea, but the wrong way of implementing it.
There are a few papers actually that describe how to get faster results and more economic sessions by instructing the LLM how to compress its thinking (“CCoT” is a paper that I remember, compressed chain of thought). It basically tells the model to think like “a -> b”. There’s loss in quality, though, but not too much.
For the more important sessions, I like to have it revise the plan with a generic prompt (e.g. "perform a sanity check") just so that it can take another pass on the beginning portion of the plan with the benefit of additional context that it had reasoned out by the end of the first draft.
Is this true? Non-reasoning LLMs are autoregressive. Reasoning LLMs can emit thousands of reasoning tokens before "line 1" where they write the answer.
A "reasoning" LLM is just an LLM that's been instructed or trained to start every response with some text wrapped in <BEGIN_REASONING></END_REASONING> or similar. The UI may show or obscure this part. Then when the model decides to give its "real" response, it has all that reasoning text in its context window, helping it generate a better answer.
LLMs are autoregressive (filling in the completion of what came before), so you'd better have thinking mode on or the "reasoning" is pure confirmation bias seeded by the answer that gets locked in via the first output tokens.