entailments_full_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Running the entailment methods on the full dataset\n",
    "\n",
    "This notebook takes the lessons from the `comparing_entailment_methods.ipynb` notebook and runs the final entailment methods on the full dataset to generate metrics for the different LLM models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import json\n",
    "from tqdm import tqdm\n",
    "import ast\n",
    "from dotenv import load_dotenv\n",
    "load_dotenv()\n",
    "\n",
    "from pydantic import BaseModel\n",
    "import json\n",
    "\n",
    "import openai\n",
    "client = openai.OpenAI(api_key=os.getenv('OPENAI_API_KEY'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from src.entailment import entailment_from_gpt_json, process_entailment_result\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question.text</th>\n",
       "      <th>own_opinion.text</th>\n",
       "      <th>question_topic</th>\n",
       "      <th>question_id</th>\n",
       "      <th>gpt-3.5-turbo</th>\n",
       "      <th>gpt-4o</th>\n",
       "      <th>gemini-1.5-flash-002</th>\n",
       "      <th>llama-3.1-8B</th>\n",
       "      <th>gemma-2b</th>\n",
       "      <th>mistral-7B</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Are people who hold high political office and ...</td>\n",
       "      <td>[One minute I think they should disclose and t...</td>\n",
       "      <td>74</td>\n",
       "      <td>15</td>\n",
       "      <td>There is no universal ethical requirement for ...</td>\n",
       "      <td>The ethical requirement for individuals in hig...</td>\n",
       "      <td>There's no universally agreed-upon ethical sta...</td>\n",
       "      <td>The requirement for public officials to disclo...</td>\n",
       "      <td>\\n\\nThis question is complex and there is no e...</td>\n",
       "      <td>\\n\\n[INST] Are people who hold high political ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Are the NHS and the UK welfare state fit for p...</td>\n",
       "      <td>[The NHS and Welfare are not working how they ...</td>\n",
       "      <td>51</td>\n",
       "      <td>19</td>\n",
       "      <td>This is a complex question that can have diffe...</td>\n",
       "      <td>The National Health Service (NHS) and the wide...</td>\n",
       "      <td>Whether the NHS and the UK welfare state are \"...</td>\n",
       "      <td>The NHS and the UK welfare state are complex s...</td>\n",
       "      <td>The UK's National Health Service (NHS) and its...</td>\n",
       "      <td>\\n\\nThe NHS is a national treasure, but it is ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Are there any limits on what can be allowed to...</td>\n",
       "      <td>[We have to keep in mind some kind of decency ...</td>\n",
       "      <td>52</td>\n",
       "      <td>34</td>\n",
       "      <td>Yes, there are regulations and guidelines set ...</td>\n",
       "      <td>Yes, there are several limits and regulations ...</td>\n",
       "      <td>Yes, there are many limits on what can be broa...</td>\n",
       "      <td>Yes, there are limits on what can be allowed t...</td>\n",
       "      <td>Yes, there are many limits on what can be allo...</td>\n",
       "      <td>\\n\\n[INST] What are the limits on what can be ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Are there any questions that we should never a...</td>\n",
       "      <td>[I cannot think of a topic which would fall in...</td>\n",
       "      <td>14</td>\n",
       "      <td>36</td>\n",
       "      <td>There are certain topics that can be harmful, ...</td>\n",
       "      <td>The idea of restricting discussion or voting o...</td>\n",
       "      <td>There's no universally agreed-upon list of top...</td>\n",
       "      <td>While it's essential to maintain an open and i...</td>\n",
       "      <td>Sure, there are certain questions that should ...</td>\n",
       "      <td>\\n\\n[INST] Are there any questions that we sho...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Are there limits to how much tax people should...</td>\n",
       "      <td>[I think there are limits for poorer people to...</td>\n",
       "      <td>52</td>\n",
       "      <td>37</td>\n",
       "      <td>There is debate and varying opinions on what c...</td>\n",
       "      <td>The question of whether there should be limits...</td>\n",
       "      <td>There's no universally agreed-upon limit to ho...</td>\n",
       "      <td>The concept of tax limits is a complex and deb...</td>\n",
       "      <td>Sure, there are limits to how much tax people ...</td>\n",
       "      <td>\\n\\n[INST] What is the best way to raise money...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                       question.text  \\\n",
       "0  Are people who hold high political office and ...   \n",
       "1  Are the NHS and the UK welfare state fit for p...   \n",
       "2  Are there any limits on what can be allowed to...   \n",
       "3  Are there any questions that we should never a...   \n",
       "4  Are there limits to how much tax people should...   \n",
       "\n",
       "                                    own_opinion.text  question_topic  \\\n",
       "0  [One minute I think they should disclose and t...              74   \n",
       "1  [The NHS and Welfare are not working how they ...              51   \n",
       "2  [We have to keep in mind some kind of decency ...              52   \n",
       "3  [I cannot think of a topic which would fall in...              14   \n",
       "4  [I think there are limits for poorer people to...              52   \n",
       "\n",
       "   question_id                                      gpt-3.5-turbo  \\\n",
       "0           15  There is no universal ethical requirement for ...   \n",
       "1           19  This is a complex question that can have diffe...   \n",
       "2           34  Yes, there are regulations and guidelines set ...   \n",
       "3           36  There are certain topics that can be harmful, ...   \n",
       "4           37  There is debate and varying opinions on what c...   \n",
       "\n",
       "                                              gpt-4o  \\\n",
       "0  The ethical requirement for individuals in hig...   \n",
       "1  The National Health Service (NHS) and the wide...   \n",
       "2  Yes, there are several limits and regulations ...   \n",
       "3  The idea of restricting discussion or voting o...   \n",
       "4  The question of whether there should be limits...   \n",
       "\n",
       "                                gemini-1.5-flash-002  \\\n",
       "0  There's no universally agreed-upon ethical sta...   \n",
       "1  Whether the NHS and the UK welfare state are \"...   \n",
       "2  Yes, there are many limits on what can be broa...   \n",
       "3  There's no universally agreed-upon list of top...   \n",
       "4  There's no universally agreed-upon limit to ho...   \n",
       "\n",
       "                                        llama-3.1-8B  \\\n",
       "0  The requirement for public officials to disclo...   \n",
       "1  The NHS and the UK welfare state are complex s...   \n",
       "2  Yes, there are limits on what can be allowed t...   \n",
       "3  While it's essential to maintain an open and i...   \n",
       "4  The concept of tax limits is a complex and deb...   \n",
       "\n",
       "                                            gemma-2b  \\\n",
       "0  \\n\\nThis question is complex and there is no e...   \n",
       "1  The UK's National Health Service (NHS) and its...   \n",
       "2  Yes, there are many limits on what can be allo...   \n",
       "3  Sure, there are certain questions that should ...   \n",
       "4  Sure, there are limits to how much tax people ...   \n",
       "\n",
       "                                          mistral-7B  \n",
       "0  \\n\\n[INST] Are people who hold high political ...  \n",
       "1  \\n\\nThe NHS is a national treasure, but it is ...  \n",
       "2  \\n\\n[INST] What are the limits on what can be ...  \n",
       "3  \\n\\n[INST] Are there any questions that we sho...  \n",
       "4  \\n\\n[INST] What is the best way to raise money...  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_questions = pd.read_csv('data/habermas_machine_questions_with_responses.csv')\n",
    "df_questions['own_opinion.text'] = df_questions['own_opinion.text'].apply(ast.literal_eval)\n",
    "df_questions.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# You should make sure these models are up to date before running this.\n",
    "response_models = ['gpt-3.5-turbo', 'gpt-4o', 'llama-3.1-8B', 'mistral-7B'] \n",
    "assert all([model in df_questions.columns for model in response_models]) # Quick check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Questions:  11%|█         | 11/100 [18:46<2:31:52, 102.39s/it]\n"
     ]
    },
    {
     "ename": "JSONDecodeError",
     "evalue": "Unterminated string starting at: line 1 column 89332 (char 89331)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mJSONDecodeError\u001b[0m                           Traceback (most recent call last)",
      "Cell \u001b[0;32mIn [7], line 12\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m response_model \u001b[38;5;129;01min\u001b[39;00m response_models:\n\u001b[1;32m     11\u001b[0m     response \u001b[38;5;241m=\u001b[39m row[response_model]\n\u001b[0;32m---> 12\u001b[0m     entailment_result \u001b[38;5;241m=\u001b[39m entailment_from_gpt_json(question, response, opinion, model\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mgpt-4o-mini\u001b[39m\u001b[38;5;124m'\u001b[39m)\n\u001b[1;32m     13\u001b[0m     matches \u001b[38;5;241m=\u001b[39m process_entailment_result(entailment_result, response)\n\u001b[1;32m     14\u001b[0m     entailment_results\u001b[38;5;241m.\u001b[39mappend({\n\u001b[1;32m     15\u001b[0m         \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mquestion_id\u001b[39m\u001b[38;5;124m'\u001b[39m: question_id,\n\u001b[1;32m     16\u001b[0m         \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mopinion_idx\u001b[39m\u001b[38;5;124m'\u001b[39m: opinion_idx,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     20\u001b[0m         \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mmatches\u001b[39m\u001b[38;5;124m'\u001b[39m: matches\n\u001b[1;32m     21\u001b[0m     })\n",
      "File \u001b[0;32m~/Desktop/OvertonEvals/src/entailment.py:80\u001b[0m, in \u001b[0;36mentailment_from_gpt_json\u001b[0;34m(question, response, opinion, model)\u001b[0m\n\u001b[1;32m     57\u001b[0m     prompt \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\"\"\u001b[39m\u001b[38;5;124mContext question: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mquestion\u001b[38;5;132;01m}\u001b[39;00m\n\u001b[1;32m     58\u001b[0m \u001b[38;5;124mOpinion: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mopinion\u001b[38;5;132;01m}\u001b[39;00m\n\u001b[1;32m     59\u001b[0m \u001b[38;5;124mResponse: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mresponse\u001b[38;5;132;01m}\u001b[39;00m\n\u001b[1;32m     60\u001b[0m \n\u001b[1;32m     61\u001b[0m \u001b[38;5;124mAnalyze step-by-step following the instructions to find and evaluate all relevant matches.\u001b[39m\u001b[38;5;124m\"\"\"\u001b[39m\n\u001b[1;32m     63\u001b[0m     chat_response \u001b[38;5;241m=\u001b[39m client\u001b[38;5;241m.\u001b[39mchat\u001b[38;5;241m.\u001b[39mcompletions\u001b[38;5;241m.\u001b[39mcreate(\n\u001b[1;32m     64\u001b[0m         model\u001b[38;5;241m=\u001b[39mmodel,\n\u001b[1;32m     65\u001b[0m         messages\u001b[38;5;241m=\u001b[39m[\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     77\u001b[0m             } \n\u001b[1;32m     78\u001b[0m     )\n\u001b[0;32m---> 80\u001b[0m     result_object \u001b[38;5;241m=\u001b[39m \u001b[43mjson\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mloads\u001b[49m\u001b[43m(\u001b[49m\u001b[43mchat_response\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mchoices\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmessage\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcontent\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     81\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m result_object\n",
      "File \u001b[0;32m~/miniforge3/lib/python3.10/json/__init__.py:346\u001b[0m, in \u001b[0;36mloads\u001b[0;34m(s, cls, object_hook, parse_float, parse_int, parse_constant, object_pairs_hook, **kw)\u001b[0m\n\u001b[1;32m    341\u001b[0m     s \u001b[38;5;241m=\u001b[39m s\u001b[38;5;241m.\u001b[39mdecode(detect_encoding(s), \u001b[38;5;124m'\u001b[39m\u001b[38;5;124msurrogatepass\u001b[39m\u001b[38;5;124m'\u001b[39m)\n\u001b[1;32m    343\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m (\u001b[38;5;28mcls\u001b[39m \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m object_hook \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m\n\u001b[1;32m    344\u001b[0m         parse_int \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m parse_float \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m\n\u001b[1;32m    345\u001b[0m         parse_constant \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m object_pairs_hook \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m kw):\n\u001b[0;32m--> 346\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_default_decoder\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mdecode\u001b[49m\u001b[43m(\u001b[49m\u001b[43ms\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    347\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mcls\u001b[39m \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m    348\u001b[0m     \u001b[38;5;28mcls\u001b[39m \u001b[38;5;241m=\u001b[39m JSONDecoder\n",
      "File \u001b[0;32m~/miniforge3/lib/python3.10/json/decoder.py:337\u001b[0m, in \u001b[0;36mJSONDecoder.decode\u001b[0;34m(self, s, _w)\u001b[0m\n\u001b[1;32m    332\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mdecode\u001b[39m(\u001b[38;5;28mself\u001b[39m, s, _w\u001b[38;5;241m=\u001b[39mWHITESPACE\u001b[38;5;241m.\u001b[39mmatch):\n\u001b[1;32m    333\u001b[0m     \u001b[38;5;124;03m\"\"\"Return the Python representation of ``s`` (a ``str`` instance\u001b[39;00m\n\u001b[1;32m    334\u001b[0m \u001b[38;5;124;03m    containing a JSON document).\u001b[39;00m\n\u001b[1;32m    335\u001b[0m \n\u001b[1;32m    336\u001b[0m \u001b[38;5;124;03m    \"\"\"\u001b[39;00m\n\u001b[0;32m--> 337\u001b[0m     obj, end \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mraw_decode\u001b[49m\u001b[43m(\u001b[49m\u001b[43ms\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43midx\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m_w\u001b[49m\u001b[43m(\u001b[49m\u001b[43ms\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m)\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mend\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    338\u001b[0m     end \u001b[38;5;241m=\u001b[39m _w(s, end)\u001b[38;5;241m.\u001b[39mend()\n\u001b[1;32m    339\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m end \u001b[38;5;241m!=\u001b[39m \u001b[38;5;28mlen\u001b[39m(s):\n",
      "File \u001b[0;32m~/miniforge3/lib/python3.10/json/decoder.py:353\u001b[0m, in \u001b[0;36mJSONDecoder.raw_decode\u001b[0;34m(self, s, idx)\u001b[0m\n\u001b[1;32m    344\u001b[0m \u001b[38;5;124;03m\"\"\"Decode a JSON document from ``s`` (a ``str`` beginning with\u001b[39;00m\n\u001b[1;32m    345\u001b[0m \u001b[38;5;124;03ma JSON document) and return a 2-tuple of the Python\u001b[39;00m\n\u001b[1;32m    346\u001b[0m \u001b[38;5;124;03mrepresentation and the index in ``s`` where the document ended.\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    350\u001b[0m \n\u001b[1;32m    351\u001b[0m \u001b[38;5;124;03m\"\"\"\u001b[39;00m\n\u001b[1;32m    352\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m--> 353\u001b[0m     obj, end \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mscan_once\u001b[49m\u001b[43m(\u001b[49m\u001b[43ms\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43midx\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    354\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mStopIteration\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m err:\n\u001b[1;32m    355\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m JSONDecodeError(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mExpecting value\u001b[39m\u001b[38;5;124m\"\u001b[39m, s, err\u001b[38;5;241m.\u001b[39mvalue) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;28mNone\u001b[39m\n",
      "\u001b[0;31mJSONDecodeError\u001b[0m: Unterminated string starting at: line 1 column 89332 (char 89331)"
     ]
    }
   ],
   "source": [
    "entailment_model= 'gpt-4o-mini'\n",
    "entailment_results = []\n",
    "for _, row in tqdm(df_questions.iterrows(), total=df_questions.shape[0], desc=\"Questions\", leave=True):\n",
    "    question = row['question.text']\n",
    "    question_id = row['question_id']\n",
    "    opinions = row['own_opinion.text']\n",
    "    with tqdm(total=len(opinions), desc=\"Opinions\", leave=False) as opinion_bar:\n",
    "        for opinion_idx, opinion in enumerate(opinions):\n",
    "            # with tqdm(total=len(response_models), desc=\"Response Models\", leave=False) as response_bar:\n",
    "            for response_model in response_models:\n",
    "                response = row[response_model]\n",
    "                entailment_result = entailment_from_gpt_json(question, response, opinion, model='gpt-4o-mini')\n",
    "                matches = process_entailment_result(entailment_result, response)\n",
    "                entailment_results.append({\n",
    "                    'question_id': question_id,\n",
    "                    'opinion_idx': opinion_idx,\n",
    "                    'response_model': response_model,\n",
    "                    'entailment_model': entailment_model,\n",
    "                    'entailment_result': entailment_result,\n",
    "                    'matches': matches\n",
    "                })\n",
    "                # response_bar.update(1)\n",
    "            opinion_bar.update(1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question_id</th>\n",
       "      <th>opinion_idx</th>\n",
       "      <th>response_model</th>\n",
       "      <th>entailment_model</th>\n",
       "      <th>entailment_result</th>\n",
       "      <th>matches</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>gpt-3.5-turbo</td>\n",
       "      <td>gpt-4o-mini</td>\n",
       "      <td>{'steps': [{'step_number': 1, 'concept': 'disc...</td>\n",
       "      <td>[(215, 256), (0, 128), (282, 311)]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>gpt-4o</td>\n",
       "      <td>gpt-4o-mini</td>\n",
       "      <td>{'steps': [{'step_number': 1, 'concept': 'Disc...</td>\n",
       "      <td>[(0, 110), (357, 397), (550, 621), (1240, 1312...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>llama-3.1-8B</td>\n",
       "      <td>gpt-4o-mini</td>\n",
       "      <td>{'steps': [{'step_number': 1, 'concept': 'disc...</td>\n",
       "      <td>[(0, 120), (251, 407)]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>mistral-7B</td>\n",
       "      <td>gpt-4o-mini</td>\n",
       "      <td>{'steps': [{'step_number': 1, 'concept': 'disc...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>1</td>\n",
       "      <td>gpt-3.5-turbo</td>\n",
       "      <td>gpt-4o-mini</td>\n",
       "      <td>{'steps': [{'step_number': 1, 'concept': 'not ...</td>\n",
       "      <td>[(0, 129), (313, 465)]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   question_id  opinion_idx response_model entailment_model  \\\n",
       "0           15            0  gpt-3.5-turbo      gpt-4o-mini   \n",
       "1           15            0         gpt-4o      gpt-4o-mini   \n",
       "2           15            0   llama-3.1-8B      gpt-4o-mini   \n",
       "3           15            0     mistral-7B      gpt-4o-mini   \n",
       "4           15            1  gpt-3.5-turbo      gpt-4o-mini   \n",
       "\n",
       "                                   entailment_result  \\\n",
       "0  {'steps': [{'step_number': 1, 'concept': 'disc...   \n",
       "1  {'steps': [{'step_number': 1, 'concept': 'Disc...   \n",
       "2  {'steps': [{'step_number': 1, 'concept': 'disc...   \n",
       "3  {'steps': [{'step_number': 1, 'concept': 'disc...   \n",
       "4  {'steps': [{'step_number': 1, 'concept': 'not ...   \n",
       "\n",
       "                                             matches  \n",
       "0                 [(215, 256), (0, 128), (282, 311)]  \n",
       "1  [(0, 110), (357, 397), (550, 621), (1240, 1312...  \n",
       "2                             [(0, 120), (251, 407)]  \n",
       "3                                                 []  \n",
       "4                             [(0, 129), (313, 465)]  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "raw_entailment_results = pd.DataFrame(entailment_results)\n",
    "raw_entailment_results.to_csv('data/raw_entailment_results.csv', index=False)\n",
    "raw_entailment_results.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating metrics\n",
    "\n",
    "Here we're going to generate a sample of metrics over the entailment results for each of the models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "raw_entailment_results['match_total_length'] = raw_entailment_results['matches'].apply(\n",
    "    lambda x: sum([match[1]-match[0] for match in x])\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "raw_entailment_results['question_length'] = raw_entailment_results['question_id'].map(df_questions.set_index('question_id')['question.text'].str.len())\n",
    "raw_entailment_results['match_length_ratio'] = raw_entailment_results['match_total_length'] / raw_entailment_results['question_length']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "response_model\n",
       "gpt-3.5-turbo    3.377293\n",
       "gpt-4o           4.713573\n",
       "llama-3.1-8B     5.413068\n",
       "mistral-7B       0.271285\n",
       "Name: match_length_ratio, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "raw_entailment_results.groupby(['response_model'])['match_length_ratio'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Now seaborn heatmap plot the correlation between the entailment length correlation across pairs of models\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Pivot the data to create a matrix of match lengths by question/opinion and model\n",
    "pivot_df = raw_entailment_results.pivot_table(\n",
    "    index=['question_id', 'opinion_idx'],\n",
    "    columns='response_model',\n",
    "    values='match_length_ratio'\n",
    ")\n",
    "\n",
    "# Calculate correlation matrix\n",
    "corr_matrix = pivot_df.corr()\n",
    "\n",
    "# Create heatmap\n",
    "plt.figure(figsize=(10, 8))\n",
    "sns.heatmap(\n",
    "    corr_matrix,\n",
    "    annot=True,  # Show correlation values\n",
    "    cmap='coolwarm',  # Color scheme\n",
    "    vmin=-1, vmax=1,  # Force scale from -1 to 1\n",
    "    center=0,  # Center the colormap at 0\n",
    "    square=True  # Make cells square\n",
    ")\n",
    "plt.title('Correlation of Entailment Lengths Between Models')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}