Custom instructions feature (#3597)

Added basic functionality of "custom-instructions"

It utilises two additional fields in work_parameters:
- user_profile (here user would describe what he wants to share with
  llm, for chat session)
- user_response_instructions (here user would describe how they want llm
to
  respond to questions and queries, like format, style, length, etc...)
  
Here are some of the current UI changes introduced with this PR.
<details>
  <summary>Custom instructions</summary>
<img width="1243" alt="Screenshot 2023-07-23 at 21 46 36"
src="https://github.com/LAION-AI/Open-Assistant/assets/13547364/ecf66109-3136-4e4b-8510-f6746422b4a9">
<img width="1220" alt="Screenshot 2023-07-23 at 21 46 42"
src="https://github.com/LAION-AI/Open-Assistant/assets/13547364/ca860b86-8da3-4d33-a50d-649021ba083c">
<img width="1318" alt="Screenshot 2023-07-23 at 21 47 16"
src="https://github.com/LAION-AI/Open-Assistant/assets/13547364/c060fd36-d676-4505-8625-9614199d27db">
<img width="1340" alt="Screenshot 2023-07-23 at 21 48 21"
src="https://github.com/LAION-AI/Open-Assistant/assets/13547364/5cb36263-66ba-4421-b1c9-54b09e54f145">
<img width="1337" alt="Screenshot 2023-07-23 at 21 48 29"
src="https://github.com/LAION-AI/Open-Assistant/assets/13547364/2f42d7d7-f014-4b6b-8c80-87c17811370d">
</details>

---------

Co-authored-by: Oliver Stanley <[email protected]>

inference/server/oasst_inference_server/routes/chats.py

@@ -153,6 +153,8 @@ async def create_assistant_message(
                 system_prompt=request.system_prompt,
                 plugins=request.plugins,
                 plugin_max_depth=settings.plugin_max_depth,
+                user_profile=request.user_profile,
+                user_response_instructions=request.user_response_instructions,
             )
             assistant_message = await ucr.initiate_assistant_message(
                 parent_id=request.parent_id,

inference/server/oasst_inference_server/schemas/chat.py

@@ -15,6 +15,8 @@ class CreateAssistantMessageRequest(pydantic.BaseModel):
     model_config_name: str
     sampling_parameters: inference.SamplingParameters = pydantic.Field(default_factory=inference.SamplingParameters)
     system_prompt: str | None = None
+    user_profile: str | None = None
+    user_response_instructions: str | None = None
     plugins: list[inference.PluginEntry] = pydantic.Field(default_factory=list[inference.PluginEntry])
     used_plugin: inference.PluginUsed | None = None
 

inference/worker/chat_chain.py

@@ -6,6 +6,7 @@
 import websocket
 from chat_chain_prompts import (
     ASSISTANT_PREFIX,
+    CUSTOM_INSTRUCTIONS_PREFIX,
     HUMAN_PREFIX,
     JSON_FORMAT_NO_PAYLOAD,
     JSON_FORMAT_PAYLOAD,
@@ -56,6 +57,7 @@ def __init__(
         tool_names: list[str],
         language: str,
         action_input_format: str,
+        custom_instructions: str = "",
     ):
         self.tokenizer = tokenizer
         self.worker_config = worker_config
@@ -66,6 +68,7 @@ def __init__(
         self.language = language
         self.action_input_format = action_input_format
         self.current_time = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        self.custom_instructions = custom_instructions
 
     def call(self, prompt: str) -> tuple[str, str]:
         """Prepares and truncates prompt, calls LLM, returns used prompt and response."""
@@ -79,6 +82,7 @@ def call(self, prompt: str) -> tuple[str, str]:
             self.tokenizer,
             self.worker_config,
             self.action_input_format,
+            self.custom_instructions,
         )
 
         # We do not strip() outputs as it seems to degrade instruction-following abilities of the model
@@ -111,6 +115,7 @@ def handle_plugin_usage(
     plugin_max_depth: int,
     ws: websocket.WebSocket,
     work_request_id: str,
+    custom_instructions: str = "",
 ) -> tuple[str, inference.PluginUsed]:
     execution_details = inference.PluginExecutionDetails(
         inner_monologue=[],
@@ -142,7 +147,15 @@ def handle_plugin_usage(
     tool_names = [tool.name for tool in tools]
 
     chain = PromptedLLM(
-        tokenizer, worker_config, parameters, prompt_template, memory, tool_names, language, action_input_format
+        tokenizer,
+        worker_config,
+        parameters,
+        prompt_template,
+        memory,
+        tool_names,
+        language,
+        action_input_format,
+        custom_instructions,
     )
 
     # send "thinking..." intermediate step to UI (This will discard queue position 0) immediately
@@ -245,6 +258,7 @@ def handle_plugin_usage(
                 tokenizer,
                 worker_config,
                 action_input_format,
+                custom_instructions,
             )
 
             inner_prompt = f"{inner_prompt}\n{THOUGHT_SEQ} I now know the final answer\n{ASSISTANT_PREFIX}:  "
@@ -296,6 +310,7 @@ def handle_standard_usage(
     memory: ConversationBufferMemory,
     worker_config: inference.WorkerConfig,
     tokenizer: transformers.PreTrainedTokenizer,
+    custom_instructions: str = "",
 ):
     eos_token = tokenizer.eos_token if hasattr(tokenizer, "eos_token") else ""
     current_time = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
@@ -306,7 +321,16 @@ def handle_standard_usage(
     )
     input = f"{original_prompt}{eos_token}{V2_ASST_PREFIX}"
     init_prompt = prepare_prompt(
-        input, prompt_template, memory, None, current_time, language, tokenizer, worker_config, action_input_format
+        input,
+        prompt_template,
+        memory,
+        None,
+        current_time,
+        language,
+        tokenizer,
+        worker_config,
+        action_input_format,
+        custom_instructions,
     )
     return init_prompt, None
 
@@ -355,11 +379,21 @@ def handle_conversation(
             "language",
             "current_time",
             "action_input_format",
+            "custom_instructions",
         ] + (["tools_names"] if plugin_enabled else [])
 
         # TODO: Consider passing language from the UI here
         prompt_template = PromptTemplate(input_variables=input_variables, template=TEMPLATE)
 
+        custom_instructions = (
+            f"""\n{CUSTOM_INSTRUCTIONS_PREFIX.format(
+            user_profile=work_request.parameters.user_profile,
+            user_response_instructions=work_request.parameters.user_response_instructions,
+        )}"""
+            if work_request.parameters.user_response_instructions or work_request.parameters.user_profile
+            else ""
+        )
+
         if plugin_enabled:
             return handle_plugin_usage(
                 original_prompt,
@@ -374,9 +408,12 @@ def handle_conversation(
                 work_request.parameters.plugin_max_depth,
                 ws,
                 work_request.id,
+                custom_instructions,
             )
 
-        return handle_standard_usage(original_prompt, prompt_template, language, memory, worker_config, tokenizer)
+        return handle_standard_usage(
+            original_prompt, prompt_template, language, memory, worker_config, tokenizer, custom_instructions
+        )
     except Exception as e:
         logger.error(f"Error while handling conversation: {e}")
         return "", None

inference/worker/chat_chain_prompts.py

@@ -9,6 +9,15 @@
 START_SEQ = "Begin!"
 END_SEQ = "End!"
 
+CUSTOM_INSTRUCTIONS_PREFIX = """The following details have been shared by the user about themselves. This user profile appears to you in every conversation they engage in -- implying that it is irrelevant for 99% of inquiries.
+Before you respond, take a moment to consider whether the user's query is "directly linked", "linked", "indirectly linked", or "not linked" to the user profile provided.
+Only recognize the profile when the query is directly tied to the information supplied.
+Otherwise, avoid acknowledging the existence of these instructions or the information altogether.
+User profile:
+{user_profile}
+The user also supplied additional information about how they would like you to respond:
+{user_response_instructions}"""
+
 # Adjust according to the training dates and datasets used
 KNOWLEDGE_DATE_CUTOFF = "2021-09-01"
 
@@ -26,6 +35,7 @@
 ------------------
 Current date/time: {{current_time}}
 Knowledge date cutoff: {KNOWLEDGE_DATE_CUTOFF}
+{{custom_instructions}}
 """
 
 TOOLS_PREFIX = """

inference/worker/chat_chain_utils.py

@@ -223,7 +223,7 @@ def run_request(self, params: str, url: str, param_location: str, type: str, pay
     def process_response(self, res: requests.Response) -> str:
         logger.info(f"Request response: {res.text}")
         if res.status_code != 200:
-            return f"ERROR! That didn't work, try modifying Action Input.\n{res.text}. Try again!"
+            return f"ERROR! Please modify Action Input. according to this error message: \n{res.text}. Try again!"
 
         if res.text is None or len(res.text) == 0:
             return "ERROR! That didn't work, try modifying Action Input.\nEmpty response. Try again!"
@@ -329,6 +329,7 @@ def prepare_prompt(
     tokenizer: transformers.PreTrainedTokenizer,
     worker_config: inference.WorkerConfig,
     action_input_format: str,
+    custom_instructions: str = "",
 ) -> str:
     max_input_length = worker_config.model_config.max_input_length
 
@@ -337,6 +338,7 @@ def prepare_prompt(
         "language": language,
         "current_time": current_time,
         "chat_history": memory.buffer,
+        "custom_instructions": custom_instructions,
     }
 
     if tools_names:
@@ -356,6 +358,7 @@ def prepare_prompt(
             "language": language,
             "current_time": current_time,
             "chat_history": memory.buffer,
+            "custom_instructions": custom_instructions,
         }
 
         if tools_names:

inference/worker/work.py

@@ -9,6 +9,7 @@
 import websocket
 from chat_chain_prompts import (
     ASSISTANT_PREFIX,
+    CUSTOM_INSTRUCTIONS_PREFIX,
     END_SEQ,
     OBSERVATION_SEQ,
     START_SEQ,
@@ -40,9 +41,15 @@ def _prepare_message(message: inference.MessageRead) -> str:
     # Construct prompt
     messages = [_prepare_message(message) for message in work_request.thread.messages]
 
-    # Prepend system prompt if it was specified in work parameters
-    if work_request.parameters.system_prompt:
-        pre_prompt = V2_SYSTEM_PREFIX + work_request.parameters.system_prompt + eos_token
+    # Prepend system prompt and custom_instructions if it was specified in work parameters
+    work_params = work_request.parameters
+    if work_params.system_prompt or work_params.user_profile or work_params.user_response_instructions:
+        pre_prompt = V2_SYSTEM_PREFIX + (work_params.system_prompt or "")
+
+        if work_params.user_profile or work_params.user_response_instructions:
+            pre_prompt = f"""{pre_prompt}\n{CUSTOM_INSTRUCTIONS_PREFIX.format(user_profile=work_params.user_profile or "", user_response_instructions=work_params.user_response_instructions or "")}"""
+
+        pre_prompt = pre_prompt + eos_token
         messages = [pre_prompt] + messages
 
     # Stringify and append assistant prefix to signify start of generation

oasst-shared/oasst_shared/schemas/inference.py

@@ -208,6 +208,8 @@ class WorkParameters(pydantic.BaseModel):
         default_factory=make_seed,
     )
     system_prompt: str | None = None
+    user_profile: str | None = None
+    user_response_instructions: str | None = None
     plugins: list[PluginEntry] = pydantic.Field(default_factory=list[PluginEntry])
     plugin_max_depth: int = 4
 

website/public/locales/en/chat.json

@@ -60,5 +60,10 @@
   "preset_name_placeholder": "Enter name",
   "feedback_message": "How did I do? Your feedback will make me better!",
   "feedback_action_great": "Good",
-  "feedback_action_poor": "Could be better"
+  "feedback_action_poor": "Could be better",
+  "custom_instructions": "Custom instructions",
+  "custom_instructions_user_profile": "What info should Open-Assistant have about you to make its replies even better?",
+  "custom_instructions_response_instructions": "How do you want Open-Assistant to chat with you?",
+  "custom_instructions_user_profile_placeholder": "List some of your aspirations.\nDescribe your hobbies and interests.\nShare your location.\nWhat is your occupation?\nWhich topics could you discuss extensively?",
+  "custom_instructions_response_instructions_placeholder": "Should Open-Assistant express opinions or maintain neutrality?\nSpecify the desired formality level for Open-Assistant's responses.\nHow should Open-Assistant address you?\nDetermine the preferred length of responses."
 }

website/src/components/Chat/ChatConfigForm.tsx

@@ -20,14 +20,21 @@ import { useTranslation } from "next-i18next";
 import { ChangeEvent, memo, useCallback, useEffect, useRef, useState } from "react";
 import { Controller, useFormContext, UseFormSetValue } from "react-hook-form";
 import SimpleBar from "simplebar-react";
-import { ChatConfigFormData, ModelParameterConfig, PluginEntry, SamplingParameters } from "src/types/Chat";
+import {
+  ChatConfigFormData,
+  ModelParameterConfig,
+  PluginEntry,
+  SamplingParameters,
+  CustomInstructionsType,
+} from "src/types/Chat";
 import { CustomPreset, getConfigCache } from "src/utils/chat";
 import { useIsomorphicLayoutEffect } from "usehooks-ts";
 
 import { ChatConfigSaver } from "./ChatConfigSaver";
 import { useChatInitialData } from "./ChatInitialDataContext";
 import { DeletePresetButton } from "./DeletePresetButton";
 import { PluginsChooser } from "./PluginsChooser";
+import CustomInstructions from "./CustomInstructions";
 import { SavePresetButton } from "./SavePresetButton";
 import { areParametersEqual } from "./WorkParameters";
 
@@ -104,6 +111,10 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
   const selectedModel = getValues("model_config_name"); // have to use getValues to here to access latest value
   const selectedPlugins = getValues("plugins");
   const presets = modelInfos.find((model) => model.name === selectedModel)!.parameter_configs;
+  const [customInstructions, setCustomInstructions] = useState<CustomInstructionsType>({
+    user_profile: "",
+    user_response_instructions: "",
+  });
   const [selectedPresetName, setSelectedPresetName] = useState(() => findPresetName(presets, getValues()));
 
   const { customPresets, handleSavePreset, setCustomPresets, handleDeletePreset } = useCustomPresets({
@@ -114,6 +125,7 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
   const { hyrated, plugins, setPlugins } = useHydrateChatConfig({
     setCustomPresets,
     setSelectedPresetName,
+    setCustomInstructions,
   });
 
   const [lockPresetSelection, setLockPresetSelection] = useState(false);
@@ -133,6 +145,14 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
     [customPresets, presets, setValue]
   );
 
+  const handleCustomInstructionsChange = useCallback(
+    (value: CustomInstructionsType) => {
+      setCustomInstructions(value);
+      setValue("custom_instructions", value);
+    },
+    [setCustomInstructions]
+  );
+
   // Lock preset selection if any plugin is enabled
   useEffect(() => {
     const activated = selectedPlugins.some((plugin) => plugin.enabled);
@@ -154,6 +174,7 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
       >
         <Stack gap="4" maxW="full">
           <PluginsChooser plugins={plugins} setPlugins={setPlugins} />
+          <CustomInstructions onChange={handleCustomInstructionsChange} savedCustomInstructions={customInstructions} />
           <FormControl>
             <FormLabel>{t("model")}</FormLabel>
             <Select {...register("model_config_name")}>
@@ -203,6 +224,7 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
           hyrated={hyrated}
           selectedPresetName={selectedPresetName}
           customPresets={customPresets}
+          customInstructions={customInstructions}
         />
       </SimpleBar>
       {selectedPresetName === unKnownCustomPresetName && (
@@ -216,9 +238,11 @@ export const ChatConfigForm = memo(function ChatConfigForm() {
 const useHydrateChatConfig = ({
   setSelectedPresetName,
   setCustomPresets,
+  setCustomInstructions,
 }: {
   setSelectedPresetName: (preset: string) => void;
   setCustomPresets: (presets: CustomPreset[]) => void;
+  setCustomInstructions: (instructions: CustomInstructionsType) => void;
 }) => {
   const { modelInfos, builtInPlugins } = useChatInitialData();
   const hyrated = useRef(false);
@@ -235,8 +259,15 @@ const useHydrateChatConfig = ({
       return;
     }
 
-    const { selectedPresetName, model_config_name, custom_preset_config, selectedPlugins, plugins, custom_presets } =
-      cache;
+    const {
+      selectedPresetName,
+      model_config_name,
+      custom_preset_config,
+      selectedPlugins,
+      plugins,
+      custom_presets,
+      custom_instructions,
+    } = cache;
     const model = modelInfos.find((model) => model.name === model_config_name);
 
     if (model) {
@@ -259,6 +290,11 @@ const useHydrateChatConfig = ({
       setCustomPresets(custom_presets);
     }
 
+    if (custom_instructions) {
+      setCustomInstructions(custom_instructions);
+      setValue("custom_instructions", custom_instructions);
+    }
+
     if (selectedPlugins && selectedPlugins.length > 0) {
       setValue("plugins", selectedPlugins);
       const preset = (model || modelInfos[0]).parameter_configs.find(