NFT Evaluator

[IkigaiLabsETH]

To use evaluators from the Eliza framework in the context of adding knowledge about curated NFT collections to your AI agent, you can create a custom evaluator that assesses and extracts relevant information from NFT-related conversations. Here's how you can implement this:

Custom NFT Evaluator

Create a new evaluator specifically for NFT collections:

import { Evaluator, IAgentRuntime, Memory, State } from "@ai16z/eliza";

const nftCollectionEvaluator: Evaluator = {
  evaluate: async (runtime: IAgentRuntime, message: Memory, state: State) => {
    const content = message.content.toLowerCase();
    
    // Extract relevant NFT information
    const extractedInfo = {
      mentionsCollection: content.includes("collection") || content.includes("nft"),
      mentionsFloorPrice: content.includes("floor price") || content.includes("floor"),
      mentionsVolume: content.includes("volume") || content.includes("trading volume"),
      mentionsRarity: content.includes("rare") || content.includes("rarity"),
    };

    // Update state with extracted information
    return {
      ...state,
      nftKnowledge: {
        ...state.nftKnowledge,
        ...extractedInfo,
      },
    };
  },
};

Integration with AgentRuntime

Register the NFT collection evaluator with the AgentRuntime:

runtime.registerEvaluator(nftCollectionEvaluator);

Enhancing Agent Responses

Use the evaluated information to enhance the agent's responses:

const enhanceResponse = (response: string, state: State) => {
  const { nftKnowledge } = state;

  if (nftKnowledge.mentionsCollection) {
    response += " Would you like to know more about specific NFT collections?";
  }

  if (nftKnowledge.mentionsFloorPrice) {
    response += " I can provide information on floor prices for popular collections.";
  }

  if (nftKnowledge.mentionsVolume) {
    response += " I can share recent trading volume data for NFT collections.";
  }

  if (nftKnowledge.mentionsRarity) {
    response += " I can explain rarity factors in NFT collections if you're interested.";
  }

  return response;
};

By implementing this custom evaluator, your Eliza agent will be able to assess NFT-related content in conversations, extract relevant information, and use that knowledge to provide more informed and context-aware responses about NFT collections[7][8].

Citations:
[1] https://ai16z.github.io/eliza/docs/core/evaluators/
[2] https://aclanthology.org/2024.emnlp-industry.18.pdf
[3] https://www.betterevaluation.org/frameworks-guides/evaluation-career-guide/what-career-evaluation-looks/types-roles-evaluation
[4] https://www.bee.com/es/36100.html
[5] https://www.cdc.gov/mmwr/volumes/73/rr/rr7306a1.htm
[6] https://www.youtube.com/watch?v=jksX0KwmPXs
[7] https://ai16z.github.io/eliza/docs/core/evaluators/
[8] https://ai16z.github.io/eliza/docs/core/agents/
[9] https://www.youtube.com/watch?v=nux0NkxBgC0
[10] https://www.youtube.com/watch?v=xmlsILjX23s