Unverified50% confidenceFactExact time
WeAct proposes adding capability-specific LoRA adapter selection to the agent's action space, achieving on-demand capability composition through the native tool-calling interface.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedWeAct trains multiple capability-specific LoRA adapters on a frozen backbone model, with each adapter corresponding to one capability; adapter selection is modeled as an internal action of the agent, on par with calling external tools.80% similarUnverifiedMastering RAG and tool calling are the two technologies that mark the dividing line for truly having the ability to develop practical Agents.66% similarUnverifiedCompared to a single LLM call, agents need to make dynamic decisions across multiple rounds of interaction, involving infrastructure work such as tool registration and scheduling, error retries, concurrency control, and observability, which raises engineering complexity by orders of magnitude.65% similarUnverifiedBuilding general-purpose embodied intelligent agents requires the co-evolution of model capabilities, hardware platforms, and data flywheels.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1073250API
curl https://kongchang.com/api/v1/knowledge/claims/1073250MCP
get_claim(id=1073250)