metadata
license: mit
BioHopR
Paper |
Description
We introduce BioHopR, a novel benchmark designed to evaluate multi-hop, multi-answer reasoning in structured biomedical knowledge graphs.
Built from the comprehensive PrimeKG, BioHopR includes 1-hop and 2-hop reasoning tasks that reflect real-world biomedical complexities.
Prompt
We used the below to get the response of the open source LLMs.
def generate_single(model, tokenizer, question):
q="You are an expert biomedical researcher.\n"+question+"\nJust give me the answer without any explanations.\nAnswer:\n"
inputs = tokenizer(q, return_tensors="pt", return_attention_mask=False).to(DEVICE)
response = model.generate(**inputs,
do_sample=False,
temperature=0.0,
top_p=None,
num_beams=1,
no_repeat_ngram_size=3,
eos_token_id=tokenizer.eos_token_id, # End of sequence token
pad_token_id=tokenizer.eos_token_id, # Pad token
max_new_tokens=32,
)
output = tokenizer.decode(response.squeeze()[len(inputs['input_ids'][0]):], skip_special_tokens=True)
return output
def generate_multi(model, tokenizer, question):
q="You are an expert biomedical researcher.\n"+question+"\nJust give me the answers without any explanations in a bullet-pointed list.\nAnswer:\n"
inputs = tokenizer(q, return_tensors="pt", return_attention_mask=False).to(DEVICE)
response = model.generate(**inputs,
do_sample=False,
temperature=0.0,
top_p=None,
num_beams=1,
no_repeat_ngram_size=3,
eos_token_id=tokenizer.eos_token_id, # End of sequence token
pad_token_id=tokenizer.eos_token_id, # Pad token
max_new_tokens=256,
)
output = tokenizer.decode(response.squeeze()[len(inputs['input_ids'][0]):], skip_special_tokens=True)
return output