Skip to content
anythingai.space
Agentic Systemsbeginner11 min · 10 steps

Reasoning Loops: ReAct and Beyond

The control flow of an agent — from ReAct's thought-action-observation cycle to Reflexion, Plan-and-Solve, graphs and tunable thinking effort.

continueCHAIN-OF-THOUGHT: SHOW THE STEPS, THEN ANSWERQ: 3 boxes × 12 eggs, 7 broken?3 × 12 = 3636 − 7 = 29answer: 29SELF-CONSISTENCY: SAMPLE 5 CHAINS, KEEP THE MAJORITY2929312929→ 29 (4 of 5)ProblemTree of ThoughtsPath A0.2Path B0.9Path C0.4
requesttool callresponsedataerror
1 / 10
Key ideaStep 1 of 10

Reasoning frameworks: chains, votes and trees

Chain-of-thought (CoT) reasoning writes the intermediate steps before the answer. Each step builds on the last, so the logic is visible and a wrong step can be caught instead of hidden inside a confident conclusion.

CoT is one of several reasoning frameworks. Self-consistency samples several chains and keeps the majority answer. Tree of Thoughts (ToT) branches into multiple reasoning paths, scores each one (by checking intermediate results or estimating its chance of success) and continues only the most promising.

InsightMatch the framework to the problem: a chain for sequential logic, votes for verification, a tree when you must explore alternatives. All of them only think. ReAct, next, lets the agent act between thoughts.
What changed
  • +Chain of steps
  • +Majority vote
  • +Branches scored and pruned

Lesson 2 of 15 in Agentic Systems