List your company

ResearchFormerly WizWand

QMSum

Benchmarks

TaskDatasetSOTA resultTrendResults

Query-based meeting summarization

QMSum (test)

19.63ROUGE-2

34

Next Token Prediction

QMSum

47Next Token Accuracy

32

Long-context language generation

QMSum

3.15Average Acceptance Length (τ)

25

Query-focused Summarization

QMSum (test)

38.67ROUGE-1

23

QMSum (val)

0.238ROUGE-L

17

Traceback (Prompt Injection Attacks)

QMSum

99Precision

13

Query-based Summarization

QMSum Golden v1 (test)

40.85ROUGE-1

11

Abstractive Summarization

QMSum

6.75BLEU

11

Meeting Summarization

QMSum

4.47Coherence

10

Query-focused summarization

QMSum

35.8ROUGE-1

10

Context Traceback

QMSum LongBench

99Precision

10

Synthetic Text Generation

QMSum

52Mean Embedding Similarity

10

Query-based Meeting Summarization

QMSum

14.734Overall Score

10

Document Summarization

QMSum (test)

38.9ROUGE-1

10

Long-context retrieval

QMSum LongEmbed

0.813nDCG@10

9

Document Summarization

QMSum

15.47G-mean

9

QMSum

0.3Std Dev ROUGE-1

8

Payload-splitting attack detection

QMSum

81Precision (QMSum)

6

Query-focused Meeting Summarization

QMSum 50 samples

4.88Fluency

6

QMSum (test)

4.93Fluency

5

QMSum

58.82R@10

4

Next Token Prediction

QMSum

32.82Acc (BERT-Small, Epsilon=Inf)

4

Abstractive Meeting Summarization

QMSum

1.67Coreference

4

Meeting Summarization

QMSum (all turns)

34.03ROUGE-1

4

QMSum Golden (test)

65Best Score

3