Datasets:
Tasks:
Text Generation
Modalities:
Text
Languages:
English
Size:
< 1K
Tags:
kv-cache
kv-cache-compression
llm-inference
inference-efficiency
efficient-inference
long-context
License:
protocol string | version string | description string | budget_ladder_percent list | budget_definition string | model_suite list | task_suite list | required_system_metrics list | method_metadata list | notes string |
|---|---|---|---|---|---|---|---|---|---|
Matched-Budget Evaluation (MBE) | 0.1.0 | Frozen evaluation specification for KV cache compression. Report at these exact settings so results are comparable. | [
50,
25,
12.5,
6.25
] | Fraction of the full-cache footprint retained, computed from M_KV = 2*B*T*L*H_kv*D_head*P. | [
{
"id": "meta-llama/Llama-3.1-8B-Instruct",
"attention": "GQA",
"tier": "7-8B"
},
{
"id": "Qwen/Qwen2.5-7B-Instruct",
"attention": "GQA",
"tier": "7-14B"
},
{
"id": "meta-llama/Llama-3.1-70B-Instruct",
"attention": "GQA",
"tier": ">=70B"
}
] | [
{
"name": "long_doc_qa",
"source": "LongBench / SCBench",
"measures": "retrieval"
},
{
"name": "multi_hop",
"source": "RULER",
"measures": "tracing + aggregation"
},
{
"name": "instruction_following",
"source": "IFEval",
"measures": "multi-instruction"
},
{
"name"... | [
"peak_kv_memory_gb",
"decode_throughput_tok_s",
"ttft_ms",
"max_batch_before_oom",
"hardware_tier"
] | [
"family",
"deployment_prerequisite",
"composability"
] | MBE consumes existing public benchmarks under their own licenses; it does not redistribute them. It fixes the axes (budgets, models, metrics) and the reporting card. |
Matched-Budget Evaluation (MBE) — KV Cache Compression
A standardized reporting protocol for KV cache compression in LLM inference. MBE is not a new task benchmark; it is a thin reporting layer that fixes which models, tasks, and budgets results are reported at, so that numbers from different papers become comparable.
- Manifest (
mbe_manifest.json): the frozen evaluation specification — model suite, task suite (consuming existing benchmarks: LongBench, RULER, SCBench, GSM8K, IFEval), the fixed KV-budget ladder (50 / 25 / 12.5 / 6.25 %), and the required system metrics. Evaluate at these exact settings so results line up. - Results (
cards/*.json): submitted KV Compression Cards — one method × one model, produced by the open harness under matched budgets.
Why
Published KV cache compression results are not comparable (different models, budgets, tasks, system metrics). MBE fixes the axes. See the companion survey and harness:
- Harness / protocol: https://github.com/rohithreddybc/kv-cache-compression-mbe
- Survey: "Breaking the Memory Wall: A Survey of Key-Value (KV) Cache Compression for Efficient Large Language Model (LLM) Inference" (Artificial Intelligence Review, under review).
How to contribute a result
Run the harness (run_mbe.py) on the manifest's model + budget ladder, then submit your
card JSON via PR to the GitHub repo or as a dataset PR here.
Citation
See CITATION.cff in the GitHub repository.
License
CC-BY-4.0. The manifest references third-party benchmarks under their own licenses.
- Downloads last month
- 7