diff --git a/lib/shared/mention-utils.js b/lib/shared/mention-utils.js
index 10392c934..b6372e219 100644
--- a/lib/shared/mention-utils.js
+++ b/lib/shared/mention-utils.js
@@ -1,193 +1,190 @@
 // @flow
 
 import { oldValidUsernameRegexString } from './account-utils.js';
 import SentencePrefixSearchIndex from './sentence-prefix-search-index.js';
 import { threadOtherMembers } from './thread-utils.js';
 import { stringForUserExplicit } from './user-utils.js';
 import { threadTypes } from '../types/thread-types-enum.js';
 import type {
   ThreadInfo,
   RelativeMemberInfo,
   ResolvedThreadInfo,
   ChatMentionCandidates,
 } from '../types/thread-types.js';
 import { idSchemaRegex, chatNameMaxLength } from '../utils/validation-utils.js';
 
 export type TypeaheadMatchedStrings = {
   +textBeforeAtSymbol: string,
   +query: string,
 };
 
 export type Selection = {
   +start: number,
   +end: number,
 };
 
 type MentionTypeaheadUserSuggestionItem = {
   +type: 'user',
   +userInfo: RelativeMemberInfo,
 };
 
 type MentionTypeaheadChatSuggestionItem = {
   +type: 'chat',
   +threadInfo: ResolvedThreadInfo,
 };
 
 export type MentionTypeaheadSuggestionItem =
   | MentionTypeaheadUserSuggestionItem
   | MentionTypeaheadChatSuggestionItem;
 
 export type TypeaheadTooltipActionItem<SuggestionItemType> = {
   +key: string,
   +execute: () => mixed,
   +actionButtonContent: SuggestionItemType,
 };
 
 // The simple-markdown package already breaks words out for us, and we are
 // supposed to only match when the first word of the input matches
 const markdownUserMentionRegex: RegExp = new RegExp(
   `^(@(${oldValidUsernameRegexString}))\\b`,
 );
 
 function isUserMentioned(username: string, text: string): boolean {
   return new RegExp(`\\B@${username}\\b`, 'i').test(text);
 }
 
 const userMentionsExtractionRegex = new RegExp(
   `\\B(@(${oldValidUsernameRegexString}))\\b`,
   'g',
 );
 
 const chatMentionRegexString = `([^\\\\]|^)(@\\[\\[(${idSchemaRegex}):((.{0,${chatNameMaxLength}}?)(?!\\\\).|^)\\]\\])`;
 const chatMentionRegex: RegExp = new RegExp(`^${chatMentionRegexString}`);
 const globalChatMentionRegex: RegExp = new RegExp(chatMentionRegexString, 'g');
 
 function encodeChatMentionText(text: string): string {
   return text.replace(/]/g, '\\]');
 }
 
 function decodeChatMentionText(text: string): string {
   return text.replace(/\\]/g, ']');
 }
 
 function getRawChatMention(threadInfo: ResolvedThreadInfo): string {
   return `@[[${threadInfo.id}:${encodeChatMentionText(threadInfo.uiName)}]]`;
 }
 
 function renderChatMentionsWithAltText(text: string): string {
   return text.replace(
     globalChatMentionRegex,
     (...match) => `${match[1]}@${decodeChatMentionText(match[4])}`,
   );
 }
 
 function extractUserMentionsFromText(text: string): string[] {
   const iterator = text.matchAll(userMentionsExtractionRegex);
   return [...iterator].map(matches => matches[2]);
 }
 
 function getTypeaheadRegexMatches(
   text: string,
   selection: Selection,
   regex: RegExp,
 ): null | RegExp$matchResult {
   if (
     selection.start === selection.end &&
     (selection.start === text.length || /\s/.test(text[selection.end]))
   ) {
     return text.slice(0, selection.start).match(regex);
   }
   return null;
 }
 
 function getMentionTypeaheadUserSuggestions(
   userSearchIndex: SentencePrefixSearchIndex,
   threadMembers: $ReadOnlyArray<RelativeMemberInfo>,
   viewerID: ?string,
   usernamePrefix: string,
 ): $ReadOnlyArray<MentionTypeaheadUserSuggestionItem> {
   const userIDs = userSearchIndex.getSearchResults(usernamePrefix);
   const usersInThread = threadOtherMembers(threadMembers, viewerID);
 
   return usersInThread
     .filter(user => usernamePrefix.length === 0 || userIDs.includes(user.id))
     .sort((userA, userB) =>
       stringForUserExplicit(userA).localeCompare(stringForUserExplicit(userB)),
     )
     .map(userInfo => ({ type: 'user', userInfo }));
 }
 
 function getMentionTypeaheadChatSuggestions(
   chatSearchIndex: SentencePrefixSearchIndex,
   chatMentionCandidates: ChatMentionCandidates,
   chatNamePrefix: string,
 ): $ReadOnlyArray<MentionTypeaheadChatSuggestionItem> {
   const result = [];
-  const threadIDs =
-    chatNamePrefix.length === 0
-      ? chatSearchIndex.entries
-      : chatSearchIndex.getSearchResults(chatNamePrefix);
+  const threadIDs = chatSearchIndex.getSearchResults(chatNamePrefix);
   for (const threadID of threadIDs) {
     if (!chatMentionCandidates[threadID]) {
       continue;
     }
     result.push({
       type: 'chat',
       threadInfo: chatMentionCandidates[threadID],
     });
   }
   return result;
 }
 
 function getNewTextAndSelection(
   textBeforeAtSymbol: string,
   entireText: string,
   query: string,
   suggestionText: string,
 ): {
   newText: string,
   newSelectionStart: number,
 } {
   const totalMatchLength = textBeforeAtSymbol.length + query.length + 1; // 1 for @ char
 
   let newSuffixText = entireText.slice(totalMatchLength);
   newSuffixText = (newSuffixText[0] !== ' ' ? ' ' : '') + newSuffixText;
 
   const newText = textBeforeAtSymbol + suggestionText + newSuffixText;
 
   const newSelectionStart = newText.length - newSuffixText.length + 1;
 
   return { newText, newSelectionStart };
 }
 
 function getUserMentionsCandidates(
   threadInfo: ThreadInfo,
   parentThreadInfo: ?ThreadInfo,
 ): $ReadOnlyArray<RelativeMemberInfo> {
   if (threadInfo.type !== threadTypes.SIDEBAR) {
     return threadInfo.members;
   }
   if (parentThreadInfo) {
     return parentThreadInfo.members;
   }
   // This scenario should not occur unless the user logs out while looking at a
   // sidebar. In that scenario, the Redux store may be cleared before ReactNav
   // finishes transitioning away from the previous screen
   return [];
 }
 
 export {
   markdownUserMentionRegex,
   isUserMentioned,
   extractUserMentionsFromText,
   getMentionTypeaheadUserSuggestions,
   getMentionTypeaheadChatSuggestions,
   getNewTextAndSelection,
   getTypeaheadRegexMatches,
   getUserMentionsCandidates,
   chatMentionRegex,
   encodeChatMentionText,
   decodeChatMentionText,
   getRawChatMention,
   renderChatMentionsWithAltText,
 };
diff --git a/lib/shared/sentence-prefix-search-index.js b/lib/shared/sentence-prefix-search-index.js
index bc85bd7cf..954aa0545 100644
--- a/lib/shared/sentence-prefix-search-index.js
+++ b/lib/shared/sentence-prefix-search-index.js
@@ -1,33 +1,29 @@
 // @flow
 
 import Tokenizer from 'tokenize-text';
 
 import SearchIndex from './search-index.js';
 
 // defaultTokenize used in SearchIndex splits on punctuation
 // We use this alternative because we only want to split on whitespace
 const tokenize = new Tokenizer().re(/\S+/);
 
 class SentencePrefixSearchIndex extends SearchIndex {
-  entries: Set<string>;
-
   constructor() {
     super(tokenize);
-    this.entries = new Set();
   }
 
   addEntry(id: string, rawText: string) {
     const keywords = this.tokenize(rawText);
     for (const keyword of keywords) {
       const value = rawText.slice(keyword.index).toLowerCase();
       this.addAllPrefixes(id, value);
     }
-    this.entries.add(id);
   }
 
   getSearchResults(query: string): string[] {
     return this.radixTree.getAllMatchingPrefix(query.toLowerCase());
   }
 }
 
 export default SentencePrefixSearchIndex;