| 1 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | HITL | What is a primary use case for "Human-in-the-Loop" (HITL) mechanisms? | Easy | 1 | A | Approval of high-stakes actions (e.g., money transfers, data deletion). | Increasing the speed of the agent's responses. | Reducing the number of tokens used by the LLM. | Replacing the LLM for simple tasks. |
| 2 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Interrupts | In LangGraph, where can you set an "Interrupt" to pause execution? | Medium | 1 | B | Inside the LLM's system prompt. | In the workflow.compile() method using interrupt_before or interrupt_after. | Only at the very beginning of the graph. | In the database configuration. |
| 3 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Persistence | What is the role of a "Checkpointer" in a LangGraph application? | Easy | 1 | C | To check the code for syntax errors. | To maintain a list of active users. | To automatically persist and restore the graph state between executions. | To limit the rate of API calls. |
| 4 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Isolation | How does LangGraph isolate different user conversations when using persistence? | Medium | 1 | B | By using different Python scripts for each user. | By using a unique thread_id in the configuration. | By creating a new database for every single turn. | By clearing the memory after every message. |
| 5 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Resuming | How do you resume a graph execution after it has been paused for human approval? | Medium | 1 | C | Restarting the script from the beginning. | Calling a manual .resume() method on the graph object. | Invoking the graph again with the same thread_id and the human's input (or None). | The graph resumes automatically after a timeout. |
| 6 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | State Updates | Which method allows an administrator to manually modify the state of a paused graph? | Hard | 1 | A | app.update_state(config, values) | app.set_state(values) | app.modify_history(config) | app.write_state(values) |
| 7 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Time Travel | What does "Time Travel" refer to in LangGraph checkpointers? | Medium | 1 | B | Running the model with future data. | Inspecting or replaying the graph state from a specific past checkpoint ID. | Accelerating the LLM inference speed. | predicting the next 10 user messages. |
| 8 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | MemorySaver | What is the primary limitation of the MemorySaver checkpointer? | Easy | 1 | A | All state data is lost when the process or server is restarted. | It is too slow for use in development labs. | It requires a complex PostgreSQL setup. | It can only store one thread at a time. |
| 9 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Savers | Which checkpointer is recommended for production environments requiring high availability and scalability? | Medium | 1 | D | MemorySaver | JsonSaver | FileSaver | PostgresSaver |
| 10 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | State Updates | What is the specific purpose of the as_node parameter in the update_state method? | Hard | 1 | B | To rename a node in the graph. | To simulate that the state update was generated by a specific node (e.g., for routing). | To delete a node from the workflow. | To create a temporary branch in the graph. |
| 11 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | HITL | Why is HITL required for financial transfers? | Easy | 1 | C | Grammar check. | System speed. | Oversight for irreversible actions. | Model training. |
| 12 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Interrupts | interrupt_after pauses execution... | Medium | 1 | D | Before the first node. | After the LLM thinks. | When the API fails. | After a specific node finishes. |
| 13 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Interrupts | What happens at an interrupt point? | Hard | 1 | A | Execution stops and returns to caller. | The program crashes. | Waits for keyboard input. | History is deleted. |
| 14 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Savers | Benefit of SQLiteSaver over MemorySaver? | Medium | 1 | B | Distributed clusters. | Persistence via local file. | Native prompt execution. | Zero disk space. |
| 15 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Isolation | Every thread in a checkpointer is unique by... | Medium | 1 | C | User IP. | System prompt. | thread_id. | Timestamp. |
| 16 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | State Updates | Manual state updates can... | Hard | 1 | D | Crash the graph. | Only be done by humans. | Change LLM weights. | Skip nodes or fix stuck workflows. |
| 17 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Time Travel | To replay from a checkpoint, you need the... | Hard | 1 | A | thread_id and checkpoint_id. | Admin password. | Graph source code. | Python version. |
| 18 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Persistence | A checkpoint contains state values and... | Hard | 1 | B | Prompt templates. | The next node to execute. | User emails. | API keys. |
| 19 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Persistence | Long-term memory is often stored in... | Medium | 1 | C | The checkpointer. | System prompts. | External Vector Databases. | Python dictionaries. |
| 20 | Unit 5: Human-in-the-Loop & Persistence | Lec5 | Savers | High availability systems use... | Easy | 1 | D | MemorySaver. | FileSaver. | thread_id only. | PostgresSaver with pooling. |