Description: Question answering with an explanation given a passage
Original dataset: https://huggingface.co/datasets/drop
---
Try querying this adapter for free in Lora Land at https://predibase.com/lora-land!
The adapter_category is Explanation and the name is Question Answering Explained (drop)
---
Sample input: Given a passage, you need to accurately identify and extract relevant spans of text that answer specific questions. Provide concise and coherent responses based on the information present in the passage as well as a reasonable coherent explanation for your response. Please format your response as a JSON payload.\n### Passage: Hoping to rebound from their divisional road loss to the Packers, the Vikings played their Week 2 home opener against the Indianapolis Colts. In the first half, kicker Ryan Longwell helped Minnesota take the early lead with a 45-yard and a 27-yard field goal in the first quarter, along with a 53-yard field goal in the second quarter. In the third quarter, the Vikes increased its lead with Longwell getting a 46-yard and a 28-yard field goal. However, the Colts responded with RB Joseph Addai getting a 1-yard TD run. In the fourth quarter, Indianapolis continued to gain ground as QB Peyton Manning completed a 32-yard TD pass to WR Reggie Wayne, followed by a two-point conversion run by RB Dominic Rhodes. Later, the Colts sealed the win with kicker Adam Vinatieri nailing the game-winning 47-yard field goal.\n### Question: How many yards was Ryan Longwell's longest field goal?\n### Answer:
---
Sample output: {"answer": "53", "explanation": "Based on the passage, Ryan Longwell made a 45-yard field goal in the first quarter, a 27-yard field goal in the first quarter, and a 53-yard field goal in the second quarter. Since the question asks for the longest field goal made by Longwell, the correct answer is 53 yards."}
---
Try using this adapter yourself!
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "mistralai/Mistral-7B-v0.1"
peft_model_id = "predibase/drop_explained"
model = AutoModelForCausalLM.from_pretrained(model_id)
model.load_adapter(peft_model_id)
