Reasoning and Acting (ReAct)
AI Methodology
Overview
Developed byPrinceton University researchers
Use casecombining reasoning and acting in language agents
Integrates with
Knowledge graph stats
Claims13
Avg confidence91%
Avg freshness100%
Last updatedUpdated 2 days ago
Trust distribution
100% unverified
Governance
Not assessed
Reasoning and Acting (ReAct)
concept
Prompting paradigm that combines reasoning traces and task-specific actions for language model agents.
Compare with...published year
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| 2022 | ○Unverified | High | Fresh | 1 |
primary use case
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| combining reasoning and acting in language agents | ○Unverified | High | Fresh | 1 |
| interactive decision making tasks | ○Unverified | High | Fresh | 1 |
| question answering with external knowledge | ○Unverified | High | Fresh | 1 |
based on
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| large language models | ○Unverified | High | Fresh | 1 |
developed by
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Princeton University researchers | ○Unverified | High | Fresh | 1 |
integrates with
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| external knowledge sources | ○Unverified | High | Fresh | 1 |
| Wikipedia API | ○Unverified | Moderate | Fresh | 1 |
requires
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| text-based environment interaction | ○Unverified | Moderate | Fresh | 1 |
supports model
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| GPT-3 | ○Unverified | Moderate | Fresh | 1 |
| PaLM | ○Unverified | Moderate | Fresh | 1 |
alternative to
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Chain-of-Thought prompting | ○Unverified | Moderate | Fresh | 1 |