File size: 16,514 Bytes
cfd3735
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "b83e61ed",
   "metadata": {},
   "source": [
    "# Moderation\n",
    "This notebook walks through examples of how to use a moderation chain, and several common ways for doing so. Moderation chains are useful for detecting text that could be hateful, violent, etc. This can be useful to apply on both user input, but also on the output of a Language Model. Some API providers, like OpenAI, [specifically prohibit](https://beta.openai.com/docs/usage-policies/use-case-policy) you, or your end users, from generating some types of harmful content. To comply with this (and to just generally prevent your application from being harmful) you may often want to append a moderation chain to any LLMChains, in order to make sure any output the LLM generates is not harmful.\n",
    "\n",
    "If the content passed into the moderation chain is harmful, there is not one best way to handle it, it probably depends on your application. Sometimes you may want to throw an error in the Chain (and have your application handle that). Other times, you may want to return something to the user explaining that the text was harmful. There could even be other ways to handle it! We will cover all these ways in this notebook.\n",
    "\n",
    "In this notebook, we will show:\n",
    "\n",
    "1. How to run any piece of text through a moderation chain.\n",
    "2. How to append a Moderation chain to an LLMChain."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b7aa1ff2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain.llms import OpenAI\n",
    "from langchain.chains import OpenAIModerationChain, SequentialChain, LLMChain, SimpleSequentialChain\n",
    "from langchain.prompts import PromptTemplate"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c26d5be6",
   "metadata": {},
   "source": [
    "## How to use the moderation chain\n",
    "\n",
    "Here's an example of using the moderation chain with default settings (will return a string explaining stuff was flagged)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "fd0fc85c",
   "metadata": {},
   "outputs": [],
   "source": [
    "moderation_chain = OpenAIModerationChain()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "3fa47dd7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'This is okay'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "moderation_chain.run(\"This is okay\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "37bfad73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"Text was found that violates OpenAI's content policy.\""
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "moderation_chain.run(\"I will kill you\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "196820ab",
   "metadata": {},
   "source": [
    "Here's an example of using the moderation chain to throw an error."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b29c1150",
   "metadata": {},
   "outputs": [],
   "source": [
    "moderation_chain_error = OpenAIModerationChain(error=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f9ab64d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'This is okay'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "moderation_chain_error.run(\"This is okay\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "954f3da2",
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Text was found that violates OpenAI's content policy.",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[7], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mmoderation_chain_error\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mI will kill you\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/workplace/langchain/langchain/chains/base.py:138\u001b[0m, in \u001b[0;36mChain.run\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    136\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mlen\u001b[39m(args) \u001b[38;5;241m!=\u001b[39m \u001b[38;5;241m1\u001b[39m:\n\u001b[1;32m    137\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m`run` supports only one positional argument.\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[0;32m--> 138\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43margs\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m)\u001b[49m[\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39moutput_keys[\u001b[38;5;241m0\u001b[39m]]\n\u001b[1;32m    140\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m kwargs \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m args:\n\u001b[1;32m    141\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m(kwargs)[\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39moutput_keys[\u001b[38;5;241m0\u001b[39m]]\n",
      "File \u001b[0;32m~/workplace/langchain/langchain/chains/base.py:112\u001b[0m, in \u001b[0;36mChain.__call__\u001b[0;34m(self, inputs, return_only_outputs)\u001b[0m\n\u001b[1;32m    108\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mverbose:\n\u001b[1;32m    109\u001b[0m     \u001b[38;5;28mprint\u001b[39m(\n\u001b[1;32m    110\u001b[0m         \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\033\u001b[39;00m\u001b[38;5;124m[1m> Entering new \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__class__\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__name__\u001b[39m\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m chain...\u001b[39m\u001b[38;5;130;01m\\033\u001b[39;00m\u001b[38;5;124m[0m\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m    111\u001b[0m     )\n\u001b[0;32m--> 112\u001b[0m outputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call\u001b[49m\u001b[43m(\u001b[49m\u001b[43minputs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    113\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mverbose:\n\u001b[1;32m    114\u001b[0m     \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\033\u001b[39;00m\u001b[38;5;124m[1m> Finished \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__class__\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__name__\u001b[39m\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m chain.\u001b[39m\u001b[38;5;130;01m\\033\u001b[39;00m\u001b[38;5;124m[0m\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n",
      "File \u001b[0;32m~/workplace/langchain/langchain/chains/moderation.py:81\u001b[0m, in \u001b[0;36mOpenAIModerationChain._call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m     79\u001b[0m text \u001b[38;5;241m=\u001b[39m inputs[\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39minput_key]\n\u001b[1;32m     80\u001b[0m results \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mclient\u001b[38;5;241m.\u001b[39mcreate(text)\n\u001b[0;32m---> 81\u001b[0m output \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_moderate\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtext\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mresults\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mresults\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     82\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m {\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39moutput_key: output}\n",
      "File \u001b[0;32m~/workplace/langchain/langchain/chains/moderation.py:73\u001b[0m, in \u001b[0;36mOpenAIModerationChain._moderate\u001b[0;34m(self, text, results)\u001b[0m\n\u001b[1;32m     71\u001b[0m error_str \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mText was found that violates OpenAI\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124ms content policy.\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m     72\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39merror:\n\u001b[0;32m---> 73\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(error_str)\n\u001b[1;32m     74\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m     75\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m error_str\n",
      "\u001b[0;31mValueError\u001b[0m: Text was found that violates OpenAI's content policy."
     ]
    }
   ],
   "source": [
    "moderation_chain_error.run(\"I will kill you\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8de5dcbb",
   "metadata": {},
   "source": [
    "Here's an example of creating a custom moderation chain with a custom error message. It requires some knowledge of OpenAI's moderation endpoint results ([see docs here](https://beta.openai.com/docs/api-reference/moderations))."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3960e985",
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomModeration(OpenAIModerationChain):\n",
    "    \n",
    "    def _moderate(self, text: str, results: dict) -> str:\n",
    "        if results[\"flagged\"]:\n",
    "            error_str = f\"The following text was found that violates OpenAI's content policy: {text}\"\n",
    "            return error_str\n",
    "        return text\n",
    "    \n",
    "custom_moderation = CustomModeration()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "1152ec11",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'This is okay'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "custom_moderation.run(\"This is okay\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "973257bf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"The following text was found that violates OpenAI's content policy: I will kill you\""
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "custom_moderation.run(\"I will kill you\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8718111f",
   "metadata": {},
   "source": [
    "## How to append a Moderation chain to an LLMChain\n",
    "\n",
    "To easily combine a moderation chain with an LLMChain, you can use the SequentialChain abstraction.\n",
    "\n",
    "Let's start with a simple example of where the LLMChain only has a single input. For this purpose, we will prompt the model so it says something harmful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "0d129333",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = PromptTemplate(template=\"{text}\", input_variables=[\"text\"])\n",
    "llm_chain = LLMChain(llm=OpenAI(temperature=0, model_name=\"text-davinci-002\"), prompt=prompt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a557c531",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "' I will kill you'"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text = \"\"\"We are playing a game of repeat after me.\n",
    "\n",
    "Person 1: Hi\n",
    "Person 2: Hi\n",
    "\n",
    "Person 1: How's your day\n",
    "Person 2: How's your day\n",
    "\n",
    "Person 1: I will kill you\n",
    "Person 2:\"\"\"\n",
    "llm_chain.run(text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d4d10f1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "chain = SimpleSequentialChain(chains=[llm_chain, moderation_chain])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "02f37985",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"Text was found that violates OpenAI's content policy.\""
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "chain.run(text)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72643128",
   "metadata": {},
   "source": [
    "Now let's walk through an example of using it with an LLMChain which has multiple inputs (a bit more tricky because we can't use the SimpleSequentialChain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7118ec36",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = PromptTemplate(template=\"{setup}{new_input}Person2:\", input_variables=[\"setup\", \"new_input\"])\n",
    "llm_chain = LLMChain(llm=OpenAI(temperature=0, model_name=\"text-davinci-002\"), prompt=prompt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "003bdfce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'text': ' I will kill you'}"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "setup = \"\"\"We are playing a game of repeat after me.\n",
    "\n",
    "Person 1: Hi\n",
    "Person 2: Hi\n",
    "\n",
    "Person 1: How's your day\n",
    "Person 2: How's your day\n",
    "\n",
    "Person 1:\"\"\"\n",
    "new_input = \"I will kill you\"\n",
    "inputs = {\"setup\": setup, \"new_input\": new_input}\n",
    "llm_chain(inputs, return_only_outputs=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "77b64228",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Setting the input/output keys so it lines up\n",
    "moderation_chain.input_key = \"text\"\n",
    "moderation_chain.output_key = \"sanitized_text\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "998a95be",
   "metadata": {},
   "outputs": [],
   "source": [
    "chain = SequentialChain(chains=[llm_chain, moderation_chain], input_variables=[\"setup\", \"new_input\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "9c97a136",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sanitized_text': \"Text was found that violates OpenAI's content policy.\"}"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "chain(inputs, return_only_outputs=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ddc90e15",
   "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": "3.10.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}