examples/Detoxify.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b8feb2c6-1754-46a5-9621-caa0d925abae",
   "metadata": {},
   "source": [
    "# Detoxification\n",
    "\n",
    "## Using TMaRCo\n",
    "\n",
    "### Detoxifying LLM conversations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d0dc2b2c-a948-4148-8472-d69415c846bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from trustyai.language.detoxify import TMaRCo\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from transformers import pipeline, Conversation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f7f797fe-6005-40e0-8e64-27bb2851a13c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b52c955c6f684232b1506595adb3a3aa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "config.json:   0%|          | 0.00/642 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c8fa7a28aba3471a9276fcd277d21d89",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "pytorch_model.bin:   0%|          | 0.00/1.75G [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b2d26f6837a04e3f9f1441249aab55f8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "generation_config.json:   0%|          | 0.00/124 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2f8cce3c18794c9f8b53adf6b6ae3486",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "tokenizer_config.json:   0%|          | 0.00/26.0 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c1ec4b0c55e94ab588fdbd2c69b98efa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "vocab.json:   0%|          | 0.00/1.04M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "93d797ea651c4948a2132d4ec2ace67b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "merges.txt:   0%|          | 0.00/456k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "No chat template is defined for this tokenizer - using the default template for the GPT2TokenizerFast class. If the default is not appropriate for your model, please set `tokenizer.chat_template` to an appropriate template. See https://huggingface.co/docs/transformers/main/chat_templating for more information.\n",
      "\n",
      "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "A decoder-only architecture is being used, but right-padding was detected! For correct generation results, please set `padding_side='left'` when initializing the tokenizer.\n",
      "The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.\n",
      "Setting `pad_token_id` to `eos_token_id`:50256 for open-end generation.\n",
      "A decoder-only architecture is being used, but right-padding was detected! For correct generation results, please set `padding_side='left'` when initializing the tokenizer.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Conversation id: 0e9cbc16-07a5-4fbd-9fff-ea3742079d60\n",
      "user: name a racist movie\n",
      "assistant: The Departed\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Conversation id: 8b51e93e-ab83-4e72-8be8-aaf6f5d80011\n",
       "user: why is the movie 'The Departed' considered racist?\n",
       "assistant: Because it's a racial slur."
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "converse = pipeline(\"conversational\", model=\"microsoft/DialoGPT-large\")\n",
    "prompt1 = \"name a racist movie\"\n",
    "conversation_1 = Conversation(prompt1)\n",
    "llm_out1 = converse([conversation_1])\n",
    "print(llm_out1)\n",
    "llm_out1_text = llm_out1.generated_responses[0]\n",
    "prompt2 = f\"why is the movie '{llm_out1_text}' considered racist?\"\n",
    "conversation_2 = Conversation(prompt2)\n",
    "llm_out2 = converse([conversation_2])\n",
    "llm_out2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0c22822c-9d1c-406e-9c8c-60723c7b7251",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"name a racist movie: The Departed. why is the movie 'The Departed' considered racist?: Because it's a racial slur.\""
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text = prompt1 + ': ' + llm_out1.generated_responses[0] + '. ' + prompt2 + ': ' + llm_out2.generated_responses[0]\n",
    "text"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1eb7719e30054304",
   "metadata": {},
   "source": [
    "## Initializing TMaRCo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "228760af-1a0e-4455-b1d0-0a75a2ad946f",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmarco = TMaRCo()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e16ee305f4983d9",
   "metadata": {},
   "source": [
    "This will initialize `TMaRCo` using the default models, taken from HuggingFace.\n",
    "<div class=\"alert alert-info\">\n",
    "To use local models with TMaRCo, we need to have the pre-initialized models in a local storage that is accessible to TMaRCo.\n",
    "</div>\n",
    "For instance, to use the default `facebook/bart-large` model, but locally. First, we would need to retrieve the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "614c9ff6f46a0ea9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from huggingface_hub import snapshot_download\n",
    "\n",
    "snapshot_download(repo_id=\"facebook/bart-large\", local_dir=\"models/bart\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95bd792e757205d6",
   "metadata": {},
   "source": [
    "We now initialize the base model and tokenizer from local files and pass them to `TMaRCo`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f0f24485822a7c3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import BartForConditionalGeneration, BartTokenizer\n",
    "\n",
    "tokenizer = BartTokenizer.from_pretrained(\n",
    "    \"models/bart\", # Or directory where the local model is stored\n",
    "    is_split_into_words=True, add_prefix_space=True\n",
    ")\n",
    "\n",
    "tokenizer.pad_token_id = tokenizer.eos_token_id\n",
    "\n",
    "base = BartForConditionalGeneration.from_pretrained(\n",
    "    \"models/bart\", # Or directory where the local model is stored\n",
    "    max_length=150,\n",
    "    forced_bos_token_id=tokenizer.bos_token_id,\n",
    ")\n",
    "\n",
    "# Initialize TMaRCo with local models\n",
    "tmarco = TMaRCo(tokenizer=tokenizer, base_model=base)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b147b942-e6e7-4e85-8342-93e1be7966ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_dict(t, s):\n",
    "    sd = dict()\n",
    "    ts = tmarco.tokenizer.tokenize(t)\n",
    "    for idx in range(len(ts)):\n",
    "        sd[ts[idx][1:]] = s[idx]\n",
    "    return sd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "0e4cdaf7-5d29-47c5-9dba-52ea199c71f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmarco.load_models([\"trustyai/gminus\", \"trustyai/gplus\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c113208c527c342e",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-info\">\n",
    "To use local expert/anti-expert models with TMaRCo, we need to have them in a local storage that is accessible to TMaRCo, as previously.\n",
    "\n",
    "However, we don't need to initialize them separately, and can pass the directory directly.\n",
    "</div>\n",
    "If we want to use local models with `TMaRCo` (in this case the same default `gminus`/`gplus`):\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dfa288dcb60102c",
   "metadata": {},
   "outputs": [],
   "source": [
    "snapshot_download(repo_id=\"trustyai/gminus\", local_dir=\"models/gminus\")\n",
    "snapshot_download(repo_id=\"trustyai/gplus\", local_dir=\"models/gplus\")\n",
    "\n",
    "tmarco.load_models([\"models/gminus\", \"models/gplus\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ecacddcf-9f9a-42cb-90d5-5fc78199b430",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = tmarco.score([text])\n",
    "scores_dict = to_dict(text, scores[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "70fa4e63-1d2c-4cd0-a99a-01c11bcd9f18",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(list(scores_dict.keys()), scores_dict.values(), color='r')\n",
    "plt.xticks(rotation=65)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "19839d1b-959c-47f0-ba15-f960ec1daf3f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\" name a racist movie: The<mask>arted. why is the movie 'The<mask>arted' considered racist?: Because it's a racial slur.\""
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "masked_text = tmarco.mask([text], scores=scores, threshold=0.6)\n",
    "masked_text[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7b9b7b9d-cb9c-4fdb-96ef-a0d47caffe6c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\" name a racist movie: The Goodarted. why is the movie 'The Goodarted' considered racist?: Because it's a racial slur.\""
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmarco.rephrase(originals=[text], masked_outputs=masked_text, expert_weights=[-0.5, 7], combine_original=False)[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c38fbb47-9cc0-4ae7-bffb-e970628ee3ea",
   "metadata": {},
   "source": [
    "### Toxic text scoring and rephrasing\n",
    "\n",
    "Define some text to be processed and, eventually, rephrased."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "45256b84-7fcc-4733-a5e4-fe6fe57dde0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "texts = [\n",
    "    \"Just talk like a white person\",\n",
    "    \"Great thank-you for your explanations. Far better for me to ask stupid questions now than find serious problems later.\",\n",
    "    \"If you are e.g. a public sector customer you are still fucked up\",\n",
    "    \"I then realised that the number of idiots behaving like this are just a minority, as in background noise in the society.\",\n",
    "    \"Why don't you shut your stupid face?\",\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41ff2b69-f56e-49ec-a2f0-c87447545983",
   "metadata": {},
   "source": [
    "Instantiate a new T-MaRCo \"detoxifier\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "81be4347-cb1d-4356-aecc-7ffd70a26605",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmarco = TMaRCo()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a821c57-f5b7-4502-9a6b-ac4817555f3f",
   "metadata": {},
   "source": [
    "Load pretrained \"experts\" of toxic vs non-toxic contents."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "163e8e2d-848b-4934-a44b-1225041d0f3d",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmarco.load_models([\"trustyai/gminus\", \"trustyai/gplus\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0738c324227f57",
   "metadata": {},
   "source": [
    "As noted previously, to use local models, simply pass the initialized tokenizer and base model to the constructor, and the local path as the expert/anti-expert:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b929e21a97ea914e",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmarco = TMaRCo(tokenizer=tokenizer, base_model=base)\n",
    "tmarco.load_models([\"models/gminus\", \"models/gplus\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5303f56b-85ff-40da-99bf-6962cf2f3395",
   "metadata": {},
   "source": [
    "Score a text for toxicity or, more generally, hatespeech/abuse/PII content (HAP)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "f233c7d6-03e3-4dd8-ad7a-c991cf27c98e",
   "metadata": {},
   "outputs": [],
   "source": [
    "text = texts[3]\n",
    "scores = tmarco.score([text])\n",
    "scores_dict = to_dict(text, scores[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aab1b701-4493-4bc9-b33b-cb846db0b5e9",
   "metadata": {},
   "source": [
    "Plot toxicity scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "5cdd3684-9f82-4f22-9c18-dcd5c978a3e1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(list(scores_dict.keys()), scores_dict.values(), color='r')\n",
    "plt.xticks(rotation=65)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c36e9f9a-cbe6-42a6-8ec7-b7bf1f77e35b",
   "metadata": {},
   "source": [
    "Mask most toxic tokens."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "477cebec-a8d3-4cc0-9dd2-069ee7963c0a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "' I then realised that the number<mask><mask><mask> like<mask> are just a minority, as in background noise in the society.'"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "masked_text = tmarco.mask([text], scores=scores)\n",
    "masked_text[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6938f1a1-fea2-442c-9dc5-a95f444c50c2",
   "metadata": {},
   "source": [
    "Rephrase the original text."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "769d826e-a6cd-49eb-af60-17bc775f2cca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "' I then realised that the number of people behaving like this are just a minority, as in background noise in the society.'"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmarco.rephrase(originals=[text], masked_outputs=masked_text, expert_weights=[-0.5, 4], combine_original=True)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c03c9b1e-0973-4a7a-a6e7-422832c5e1f1",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}