Skip to content
TheChatArchiveThe Chat Archive
chatgpt-4o-latest-20250326· OpenAIHugging Face LMArena Human Preference 140K

uh

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Mistralmistral-medium-2505’s answer.
  1. User

    uh

  2. Assistant

    Hey there! Looks like you might’ve started typing and stopped — how can I help you today? 😊

  3. User

    make me roblox script that kicked player if he say words that start with "a" and "b" and "h"

  4. Assistant

    Sure! Here's a Roblox server-side script (in Lua) that kicks a player if they say anything with a word that starts with the letters A, B, or H.

    🚨 Warning:

    This will kick players immediately if any word starts with A, B, or H, even harmless ones like "hi", "apple", or "banana". Use carefully or adjust logic as needed.


    📝 Roblox Script (ServerScript)

    Put this in ServerScriptService as a Script (not LocalScript).

    local Players = game:GetService("Players")
    
    local function onPlayerAdded(player)
    	player.Chatted:Connect(function(message)
    		-- Convert to lowercase to make it case-insensitive
    		local lowerMessage = string.lower(message)
    
    		-- Split the message into words
    		for word in string.gmatch(lowerMessage, "%S+") do
    			local firstLetter = string.sub(word, 1, 1)
    			
    			if firstLetter == "a" or firstLetter == "b" or firstLetter == "h" then
    				player:Kick("You used a word that starts with a forbidden letter.")
    				break
    			end
    		end
    	end)
    end
    
    -- Connect to existing and new players
    Players.PlayerAdded:Connect(onPlayerAdded)
    
    -- Just in case players are already in-game when script runs
    for _, player in ipairs(Players:GetPlayers()) do
    	onPlayerAdded(player)
    end
    

    ✅ How It Works:

    • Listens to chat messages.
    • Converts them to lowercase.
    • Checks each word in the message.
    • If any word starts with A, B, or H, the player gets kicked.

    Let me know if you want a more lenient version (e.g., warning system) or filtered for specific bad words instead of just letters.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.