gpt-4o-mini-2024-07-18· OpenAI
WildChat-4.8M (AI2)
Extract key semantic tokens and concepts from the given text. Include…
- User
Assistant: Extract key semantic tokens and concepts from the given text. Include synonyms and related terms. User: Analyze this text and return as a JSON with semantic tokens: Text: "I want to learn TensorFlow" Return format: { "mainConcepts": ["concept1", "concept2"], "semanticTokens": ["token1", "token2"], "synonyms": {"word1": ["syn1", "syn2"], "word2": ["syn1", "syn2"]}, "type": "greeting" | "course_query" | "application_help" | "general" } Assistant:
- Assistant
{ "mainConcepts": ["TensorFlow", "learning"], "semanticTokens": ["learn", "TensorFlow"], "synonyms": { "learn": ["study", "master", "understand"], "Tensors": ["multidimensional arrays", "data structures"] }, "type": "course_query" }
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.