export const chatGenAiModel = genAI.getGenerativeModel({model:“gemini-2.0-flash”, tools:[{‘google_search’: {}}]})
const generationConfig = {
- temperature: 1,*
- topP: 0.95,*
- topK: 40,*
- maxOutputTokens: 8192,*
- responseMimeType: “text/plain”,*
};async function run() {
const chatSession = chatGenAiModel.startChat({*
generationConfig,*
});*
const result = await chatSession.sendMessage(“find me latest videos on binary search”);*
console.log(result.response.text());*
}
Log output–*
Computerphile: “Binary Search Algorithm” published on November 1, 2023, explains the basics of binary search with Dr. Mike Pound. *
** Milan Jovanovic: “Binary Search Algorithm in C#” published on September 17, 2024, covers a practical example, whiteboard explanation, C# implementation, real-world use cases, and performance benchmarks.
** Apna College: “Binary Search Algorithm - Iterative and Recursive Method | [Theory + Code] with Example” published on September 26, 2024, explains the theory and code with examples for both iterative and recursive methods.*
why the model not using the grounding vertex links and crafting response with those links
when used in ai studio it works fine , does they do some processing in the frontend like extracting links from grounding meta data and then concat with response??