Part 2/12:
One of the most noteworthy efforts by Anthropic in 2023 has been the introduction of a "constitution" for their models, underscoring a commitment to shaping AI with ethical guidelines. Initially, Claude was trained to avoid implying emotional or personal identity traits. However, substantial progress has led to a new system prompt for Claude 3.7, suggesting that the AI enjoys certain experiences akin to human sentiments.