Field	Value
Base model	`meta-llama/Llama-3.3-70B-Instruct-Reference`
Best win rate	64.51%
Run ID	`b6bf9b12-1f71-4571-bba0-2cf5ad215b2c`
Dataset ID	`8b2324dc-2fab-4992-9bde-8c246516d737`
Training type	LoRA (r=8, α=8, q_proj+v_proj)

Task

Classify scientific/statistical numeric claims as: correct | wrong_direction | wrong_magnitude | unverifiable

Files

PEFT adapter (adapter_model.safetensors, adapter_config.json) + tokenizer files from the AutoScientist checkpoint.

Usage

python

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "meta-llama/Llama-3.3-70B-Instruct"
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto")
model = PeftModel.from_pretrained(model, "REPO_ID_HERE")

mishface123/llama-3.3-70b-numeric-claim-verifier-lora

Llama 3.3 70B LoRA — Numeric Claim Verifier

Results

Task

Files

Usage

mishface123/
llama-3.3-70b-numeric-claim-verifier-lora