7249d447 · permissionBRICK/SillyTavern

     "Duplicate": "複製",
     "Persona": "使用者角色",
     "Delete": "刪除",
-    "AI Response Configuration": "設定 AI 回應",
+    "AI Response Configuration": "AI 回應設定",
     "AI Configuration panel will stay open": "上鎖 = AI 設定面板將保持開啟",
-    "clickslidertips": "點選滑桿數字可手動輸入。",
+    "clickslidertips": "點擊滑桿旁的數字以手動輸入。",
     "MAD LAB MODE ON": "瘋狂實驗室模式",
     "Documentation on sampling parameters": "取樣參數的說明文件。",
     "kobldpresets": "Kobold 預設設定檔",

     "Only enable this if your model supports context sizes greater than 8192 tokens": "僅在您的模型支援超過 8192 個符元的上下文長度時啟用此功能",
     "Max prompt cost:": "最大提示詞費用：",
     "Display the response bit by bit as it is generated.": "逐字顯示生成中的回應內容。",
-    "When this is off, responses will be displayed all at once when they are complete.": "關閉時，回應將在生成完成後一次性顯示。",
+    "When this is off, responses will be displayed all at once when they are complete.": "關閉時，回應將在生成完成後一次全部顯示。",
     "Temperature": "溫度",
     "rep.pen": "重複懲罰",
     "Rep. Pen. Range.": "重複懲罰範圍",

     "Aggressive": "積極",
     "Very aggressive": "非常積極",
     "Unlocked Context Size": "解鎖上下文長度",
-    "Unrestricted maximum value for the context slider": "不限制上下文滑桿最大值",
+    "Unrestricted maximum value for the context slider": "不限制上下文長度的最大值",
     "Context Size (tokens)": "上下文長度（符元數）",
     "Max Response Length (tokens)": "最大回應長度（符元數）",
     "Multiple swipes per generation": "每次生成多次滑動",

     "Utility Prompts": "實用提示詞",
     "Impersonation prompt": "AI 扮演提示詞",
     "Restore default prompt": "還原預設提示詞",
-    "Prompt that is used for Impersonation function": "用於 AI 模仿功能的提示詞",
+    "Prompt that is used for Impersonation function": "用於「AI 扮演使用者」功能的提示詞",
     "World Info Format Template": "世界資訊格式",
     "Restore default format": "還原預設格式",
     "Wraps activated World Info entries before inserting into the prompt.": "在插入提示詞前包裝已啟用的世界資訊條目。",

     "scenario_format_template_part_2": "來標示要插入內容的位置。",
     "Scenario Format Template": "場景格式",
     "Personality Format Template": "個性格式",
-    "Group Nudge Prompt Template": "群組推動提示詞範本",
+    "Group Nudge Prompt Template": "群組聊天格式微調",
     "Sent at the end of the group chat history to force reply from a specific character.": "在群組聊天歷史結束時發送以強制特定角色回覆",
     "New Chat": "新聊天",
     "Restore new chat prompt": "還原新聊天的提示詞",

     "Set at the beginning of the chat history to indicate that a new group chat is about to start.": "設定在聊天歷史的開頭以表明即將開始新的群組聊天",
     "New Example Chat": "新範例聊天",
     "Set at the beginning of Dialogue examples to indicate that a new example chat is about to start.": "設定在對話範例的開頭以表明即將開始新的範例聊天",
-    "Continue nudge": "繼續輔助提示詞",
+    "Continue nudge": "繼續輔助微調",
-    "Set at the end of the chat history when the continue button is pressed.": "按下繼續按鈕時設定在聊天歷史的末尾",
+    "Set at the end of the chat history when the continue button is pressed.": "按下「繼續」按鈕時，插入於聊天歷史的末尾",
     "Replace empty message": "取代空白訊息",
-    "Send this text instead of nothing when the text box is empty.": "當文字方塊為空時，發送此字串而不是空白。",
+    "Send this text instead of nothing when the text box is empty.": "當文本框為空時，發送此文本以取代空白。",
     "Seed": "種子",
-    "Set to get deterministic results. Use -1 for random seed.": "設定以獲取確定性結果。使用 -1 作為隨機種子",
+    "Set to get deterministic results. Use -1 for random seed.": "設定數值以取得可重現的結果。使用 -1 作為隨機種子",
-    "Temperature controls the randomness in token selection": "溫度控制符元選擇中的隨機性",
+    "Temperature controls the randomness in token selection": "溫度（Temperature）控制符元選擇的隨機性。\n- 低溫（<1.0）：產生更可預測且具邏輯性的文本，優先選擇機率較高的符元。\n- 高溫（>1.0）：提升創造性與輸出的多樣性，更常選擇機率較低的符元。\n將值設為 1.0 可使用原始機率。",
     "Top_K_desc": "Top K 設定可以選擇的最高符元數量。\n例如，Top K 為 20，這意味著只保留排名前 20 的符元（無論它們的機率是多樣還是有限的）。\n設定為 0 以停用。",
     "Top_P_desc": "Top P(又名核心取樣)會將所有頂級符元加總，直到達到目標百分比。\n例如，如果前兩個符元都是 25%，而 Top P 設為 0.5，那麼只有前兩個符元會被考慮。\n設定為 1.0 以停用。",
     "Typical P": "Typical P",
-    "Typical_P_desc": "Typical P 取樣根據符元偏離集合平均熵的程度進行優先排序。\n它會保留累積機率接近預設閾值(例如0.5)的符元,強調那些具有平均信息量的符元。\n設定為 1.0 以停用。",
+    "Typical_P_desc": "Typical P 取樣根據符元偏離集合平均熵的程度進行優先排序。\n它會保留累積機率接近預設閾值(例如0.5)的符元，強調那些具有平均信息量的符元。\n設定為 1.0 以停用。",
     "Min_P_desc": "Min P 設定基本最小機率。\n這個值會根據最高符元的機率進行調整。例如，如果最高符元機率為 80%,而 Min P 設為 0.1 ，那麼只有機率高於 8% 的符元會被考慮。\n設定為 0 以停用。",
     "Top_A_desc": "Top A 根據最高符元機率的平方設定符元選擇的門檻。\n例如，如果 Top A 值為 0.2，而最高符元機率為 50%，那麼低於 5%(0.2 * 0.5^2) 的符元機率就會被排除。\n設定為 0 以停用。",
     "Tail_Free_Sampling_desc": "無尾取樣(Tail-Free Sampling, TFS)會透過分析符元機率的變化率(使用導數)來尋找分佈中的低機率符元尾部。\n它會根據標準化的二階導數，保留直到某個閾值(例如0.3)的符元。\n數值越接近 0 ，表示會棄去越多符元。設定為 1.0 以停用。",

     "Logit Bias": "Logit 偏差",
     "Add": "新增",
     "Helps to ban or reenforce the usage of certain words": "有助於禁止或強化某些符元的使用",
-    "CFG Scale": "CFG 比例",
+    "CFG Scale": "CFG 縮放比例",
     "Negative Prompt": "負面提示詞",
-    "Add text here that would make the AI generate things you don't want in your outputs.": "在這裡新增文字，使 AI 生成您不希望在輸出中出現的內容。",
+    "Add text here that would make the AI generate things you don't want in your outputs.": "在此新增文字，以防止 AI 在輸出中生成您不希望出現的內容。",
-    "Used if CFG Scale is unset globally, per chat or character": "如果CFG Scale未在全域、每個聊天或角色中設定，則使用",
+    "Used if CFG Scale is unset globally, per chat or character": "若 CFG 縮放比例未被全域設定，它將作用於所有聊天或角色",
     "Mirostat Tau": "Tau",
     "Mirostat LR": "Mirostat 學習率",
     "Min Length": "最小長度",

     "Applies additional processing to the prompt before sending it to the API.": "這個選項會在將提示詞送往 API 之前，對它進行額外的處理。",
     "Verifies your API connection by sending a short test message. Be aware that you'll be credited for it!": "透過發送簡短的測試訊息來驗證您的 API 連線。請注意，您將因此獲得榮譽！",
     "Test Message": "測試訊息",
-    "Auto-connect to Last Server": "自動連線到上次伺服器",
+    "Auto-connect to Last Server": "自動連接至上次使用的伺服器",
     "Missing key": "❌ 鑰匙遺失",
     "Key saved": "✔️ 金鑰已儲存",
     "View hidden API keys": "查看隱藏的 API 金鑰",

     "Prefer Character Card Prompt": "角色卡主要提示詞優先",
     "If checked and the character card contains a jailbreak override (Post History Instruction), use that instead": "如果選中並且角色卡包含越獄覆寫（聊天歷史後指示），則使用該提示詞。",
     "Prefer Character Card Jailbreak": "角色卡越獄優先",
-    "never_resize_avatars_tooltip": "避免裁剪和調整匯入的角色圖像大小。未勾選時將會裁剪／調整大小到 512x768。",
+    "never_resize_avatars_tooltip": "避免裁剪與調整導入的角色頭像大小。未啟用此選項時，圖片將被裁剪／調整為 512x768。此設定會關閉上傳頭像時的裁剪彈出視窗。",
     "Never resize avatars": "永不調整頭像大小",
     "Show actual file names on the disk, in the characters list display only": "僅在角色列表顯示實際檔案名稱。",
     "Show avatar filenames": "顯示頭像檔案名",

     "prompt_manager_edit": "編輯",
     "prompt_manager_name": "名稱",
     "A name for this prompt.": "這個提示詞的名稱。",
-    "To whom this message will be attributed.": "此訊息將隸屬於誰。",
+    "To whom this message will be attributed.": "此訊息所屬的角色。",
     "AI Assistant": "人工智慧助手",
     "prompt_manager_position": "位置",
     "Injection position. Next to other prompts (relative) or in-chat (absolute).": "注入位置。與其他提示詞相鄰（相對位置）或在聊天中（絕對位置）。",

     "Example: http://localhost:1234/v1": "例如：http://localhost:1234/v1",
     "popup-button-crop": "裁剪",
     "(disabled when max recursion steps are used)": "（當最大遞歸步驟數使用時將停用）",
-    "0 = unlimited, 1 = scans once and doesn't recurse, 2 = scans once and recurses once, etc": "0 = 無限制，1 = 掃描一次且不遞歸，2 = 掃描一次並遞歸一次，以此類推\n（使用最小啟動設定時將停用）",
     "A greedy, brute-force algorithm used in LLM sampling to find the most likely sequence of words or tokens. It expands multiple candidate sequences at once, maintaining a fixed number (beam width) of top sequences at each step.": "一種用於 LLM 抽樣的貪婪演算法，用於尋找最可能的單詞或標記序列。該方法會同時展開多個候選序列，並在每一步中保持固定數量的頂級序列（beam width）。",
     "A multiplicative factor to expand the overall area that the nodes take up.": "節點佔用該擴充功能區域的倍數。",
     "Abort current image generation task": "終止目前的圖片生成任務",

     "Show group chat queue": "顯示群組聊天隊列",
     "Size threshold (KB)": "大小閾值（KB）",
     "Slash Command": "斜線命令",
-    "space_ slash command.": "斜線命令。",
+    "space_ slash command.": " 斜線命令。",
     "Sprite Folder Override": "表情立繪資料夾覆蓋",
     "Sprite set:": "立繪組：",
     "Show Gallery": "查看圖庫",

     "Threshold": "閾值",
     "to install 3rd party extensions.": "用於安裝第三方擴充功能。",
     "Top": "頂部",
-    "Translate text to English before classification": "在分類前將文本翻譯為英文。",
+    "Translate text to English before classification": "分類前，將訊息翻譯為英文",
     "Uncheck to hide the extensions messages in chat prompts.": "不勾選即可隱藏聊天提示詞中的擴充功能訊息。",
     "Unchecked: only entries with ❌ status can be activated.": "未勾選時：僅允許啟用狀態為 ❌ 的條目。",
     "Unified Sampling": "統一取樣（Unified Sampling）",

     "Narrate user messages": "朗讀使用者訊息",
     "Auto Generation": "自動生成",
     "Requires auto generation to be enabled.": "需要啟用自動生成功能。",
-    "Narrate by paragraphs (when streaming)": "按段落朗讀（使用「串流」傳輸時）",
+    "Narrate by paragraphs (when streaming)": "按段落朗讀（使用「串流」時）",
     "Only narrate quotes": "僅朗讀「引號」中的文字",
     "Ignore text, even quotes, inside asterisk": "忽略 *（星號）內的文字（包括「引號」）",
     "Narrate only the translated text": "僅朗讀翻譯後的文本",

     "Modules provided by your Extras API:": "由您的 Extras API 提供的模組：",
     "Not connected to the API!": "未連線到 API！",
     "ext_type_system": "這是內建的擴充功能，無法刪除，且會跟隨系統更新。",
-    "Valid": "已驗證"
+    "Valid": "已驗證",
+    "Request Model Reasoning": "請求模型推理",
+    "Global list": "全域列表",
+    "Preset-specific list": "特定預設列表",
+    "Constrains effort on reasoning for reasoning models.": "限制推理模型的推理耗費。\n目前支援的值為低、中和高。\n降低推理耗費可加快回應速度，並減少推理所使用的符元數量。",
+    "Reasoning Effort": "推理耗費",
+    "openai_reasoning_effort_low": "低",
+    "openai_reasoning_effort_medium": "中",
+    "openai_reasoning_effort_high": "高",
+    "Reasoning": "推理 Reasoning",
+    "reasoning_auto_parse": "自動解析主要內容中推理區塊，需定義且不為空的前綴與後綴欄位。",
+    "Auto-Parse": "自動解析",
+    "reasoning_auto_expand": "自動展開推理區塊。",
+    "Auto-Expand": "自動展開",
+    "reasoning_show_hidden": "顯示隱藏推理功能模型的推理時間",
+    "Show Hidden": "顯示隱藏內容",
+    "reasoning_add_to_prompts": "將現有推理區塊添加至提示詞中。若需新增推理區塊，請使用訊息編輯選單。",
+    "Add to Prompts": "添加至提示詞",
+    "reasoning_max_additions": "從最後一則訊息起算，每則提示詞中可添加的最大推理區塊數量。",
+    "Max": "最大值",
+    "Reasoning Formatting": "推理格式",
+    "reasoning_prefix": "插入於推理內容之前。",
+    "Prefix": "前綴",
+    "reasoning_suffix": "插入於推理內容之後。",
+    "Suffix": "後綴",
+    "reasoning_separator": "插入於推理內容與訊息內容之間。",
+    "Separator": "分隔符",
+    "Character details are hidden.": "角色詳情已隱藏。",
+    "Add a reasoning block": "新增推理區塊",
+    "Thought for some time": "思考了一段時間",
+    "Confirmedit": "確認",
+    "Remove reasoning": "移除推理",
+    "Cancel edit": "取消編輯",
+    "Copy reasoning": "複製推理",
+    "Edit reasoning": "編輯推理",
+    "extension_install_1": "若要從此頁面下載擴充功能，您需要安裝",
+    "extension_install_2": "已安裝。",
+    "extension_install_3": "點擊",
+    "extension_install_4": "圖示以訪問擴充功能的儲存庫，查看使用技巧。",
+    "Use the selected API from Chat Translation extension settings.": "使用擴充功能設定中，「聊天翻譯」所選的翻譯提供者（API）。",
+    "A single expression can have multiple sprites. Whenever the expression is chosen, a random sprite for this expression will be selected.": "單個同名表情可以有多張角色立繪。每次使用該表情時，會隨機擇一顯示。",
+    "Allow multiple sprites per expression": "允許單一表情使用多張立繪",
+    "If the same expression is used again, re-roll the sprite. This only applies to expressions that have multiple available sprites assigned.": "若再次使用相同的表情，將重新隨機選擇。此功能僅適用於分配了多張立繪的表情。",
+    "Re-roll if same expression is used again": "重複使用同名表情時，隨機選用其他立繪",
+    "upload_expression_request": "請輸入角色立繪名稱（不含副檔名）。",
+    "upload_expression_naming_1": "角色立繪名稱必須符合所選表情的命名規則：{{expression}}",
+    "upload_expression_naming_2": "對於多個表情，名稱必須包含表情名稱和有效的後綴，允許的分隔符為「-」或「.」。",
+    "upload_expression_replace": "點擊「取代」以取代現有表情：",
+    "ext_regex_reasoning_desc": "推理區塊內容。當「僅格式化提示詞」已勾選時，這也會影響添加至提示詞的推理內容。",
+    "Token Counter": "符元計數器",
+    "Type / paste in the box below to see the number of tokens in the text.": "在下框中輸入或貼上文字以查看符元（Token）數量。",
+    "Selected tokenizer:": "選擇的分詞器：",
+    "Input:": "輸入：",
+    "Tokens:": "符元數：",
+    "Tokenized text:": "已符元化的文字：",
+    "Token IDs:": "符元 ID：",
+    "Narrate by paragraphs (when not streaming)": "按段落朗讀（不使用「串流」時）"
+}

Merge pull request #3539 from RivelleDays/staging Update zh-tw.json