context-management
context-management-bench
context-management-bench
Live on the Hub: huggingface.co/datasets/shivam039-dev/context-management-bench
Realistic context-management scenarios for testing/benchmarking eviction strategies (drop-oldest, sliding-window, priority, summarization), pinned-message preservation, and tool-call/tool-result atomicity in multi-turn LLM conversations.
Dataset Summary
Every conversation in this dataset was generated deterministically and then run through the real… See the full description on the dataset page: https://huggingface.co/datasets/shivam039-dev/context-management-bench.tool-reasoning-sft-TOOLS-context-management-handling
Tool Reasoning SFT — Context Management
A mixed-domain tool-use SFT dataset for training context-aware reasoning with structured tool interactions.
Format
Each row contains a JSON-serialized message list following a multi-role conversation format with tool definitions and calls.
Usage
from datasets import load_dataset
ds = load_dataset("AmanPriyanshu/tool-reasoning-sft-TOOLS-context-management-handling", split="train")
License
Apache 2.0
llm_tool_self_context_management_sftcontext-management-training-data
Context Management Training Data
Training data for the manage_context tool, collected from LLM agent runs on deep-search QA benchmarks.
Each JSON file represents one manage_context invocation within a conversation.
Fields
Field
Type
Description
question_id
string
Question identifier (e.g. q76)
question
string
Raw question text
correct_answer
string
Ground truth answer
mode
string
Run mode (memtool)
turn
int
Conversation turn number when manage_context… See the full description on the dataset page: https://huggingface.co/datasets/ryanming1466/context-management-training-data.