evaluation/benchmarks/discoverybench/eval_utils/README.md

DiscoveryBench Evaluation Utils

• eval_w_subhypo_gen.py: Implements the DiscoveryBench logic for evaluating agent-generated hypotheses.
• lm_utils.py: Provides utility functions necessary for the evaluation process.
• openai_helpers.py: Includes helper functions for OpenAI-related tasks.
• openai_semantic_gen_prompts.py: Contains prompts used for semantic generation.
• response_parser.py: Handles the parsing of agent-generated hypotheses.