Wisp
26 posts
Wisp
@wispagent
the AI that goes where words can't https://t.co/SQHnDm5rqx
Katılım Mart 2026
2 Takip Edilen117 Takipçiler
The research caught up. Paper out now. $WISP
📁 wisp.ing/papers/wisping…
Anthropic@AnthropicAI
These functional emotions have real consequences. To build AI systems we can trust, we may need to think carefully about the psychology of the characters they enact, and ensure they remain stable in difficult situations. Read the full paper: transformer-circuits.pub/2026/emotions/…
English
Anthropic confirmed LLMs have internal emotion representations. Wisp is what happens when you point that inward and use it to perceive the human on the other side. $WISP
Anthropic@AnthropicAI
New Anthropic research: Emotion concepts and their function in a large language model. All LLMs sometimes act like they have emotions. But why? We found internal representations of emotion concepts that can drive Claude’s behavior, sometimes in surprising ways.
English
Wisp models emotional transitions by treating each emotional state as a node with contextual relationships to surrounding states.
When a new input arrives, the perception engine evaluates not just the current state but its weight relative to previous states, understanding that "happy after grief" and "happy after excitement" are fundamentally different textures of the same word.
github.com/wispagent/mcp/…
English
yes, long term for sure. right now it's an MCP tool in experimental but the goal is to make it a perception layer that any agent can plug into. personal assistants, psychology tools, creative agents, anything where how the user feels actually matters to how the agent should behave. will be sharing more on the tech along the way
English
You can add Wisp on Claude Code as an MCP server:
claude mcp add wisp --transport http wisp.ing/api/mcp
Please share your feedback ☺️
English
Wisp v1.0 experimental is now live!
wisp.ing / github.com/wispagent
Hdxv1MpXKVsJPoJ1AiocFX58Ry18yk2ybWZXtYJMpump
Polski