We need two baseline models from the paper:
- LLM - only an LLM for generating the next token.
- PerBrainLLM - from paper: randomly assigns the brain recordings as inputs across different prediction tasks through a permutation, and breaks connections between the stimuli and brain responses
We should compare all model metrics to these baselines for proof of actual improvement.
We need two baseline models from the paper:
We should compare all model metrics to these baselines for proof of actual improvement.