引導式測驗 · 中等 等級
Lexis+ AI 與 Westlaw AI 測驗
測試 Lexis+ AI 和 Westlaw 如何透過檢索增強生成來接地回答、什麼是接地錯誤以及如何建立自己的測試集。
問題 1 的 8
Lexis+ AI 和 Westlaw 的人工智慧研究功能都使用什麼技術來得出答案?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- Lexis+ AI 和 Westlaw 的人工智慧研究功能都使用什麼技術來得出答案?
- Answering only from what the language model memorized during training
- Retrieval-augmented generation that searches the vendor's own legal collection before the model writes an answer
- Copying answers from public web search results
- Having human editors write each answer in real time
- 在法律研究人工智慧中,什麼是誤解?
- Citing a real case for a point it does not actually support
- Citing a case that does not exist
- Failing to return any answer
- Returning results from the wrong database vendor
- 2024 年史丹佛大學的研究對 Lexis+ AI 和 Westlaw 的人工智慧輔助研究有何發現?
- Neither tool ever produced an incorrect answer
- Only Lexis+ AI produced errors
- Both gave incorrect or misgrounded answers on a meaningful share of questions, with Westlaw's tool erring more often in that test
- The tools performed identically on every question
- 該指南將哪些編輯資產與 Westlaw 而不是 Lexis 相關聯?
- Shepard's and Lexis Practical Guidance
- Protégé and Shepard's
- Lexis headnotes and Protégé
- The West Key Number System and KeyCite
- 為什麼指南建議在測試集中包含錯誤前提問題?
- Because they make the test shorter
- Because a model may accept the premise and answer as if it were true
- Because false-premise questions are required by bar rules
- Because they only work on one of the two platforms
- If the retrieval stage misses the controlling case, what does the guide say can happen?
- The system automatically refuses to answer
- The citator flags the missing case in red
- The answer can be fluent and wrong
- The model searches the open web instead
- Which company did Thomson Reuters acquire in 2023, bringing the CoCounsel assistant into its products?
- Casetext
- Lex Machina
- Clarivate
- Bloomberg Law
- Why do multi-step, agent-style research features make checking more important rather than less?
- Because agents cannot cite sources
- Because they only work on federal law
- Because they remove citator flags from results
- Because longer outputs contain more claims to verify