Salesforce AI Research·released 2023-06-271 source
XGen-7B benchmark scores: 6 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 6 benchmarks·best result 65.6 on HellaSwag (self-reported)·0 of 6 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| HellaSwag | 65.6accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
| PIQA | 51accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
| Winogrande | 29.8accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
| ARC AI2 | 21.6accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
| OpenBookQA | 20.27accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
| MMLU | 15.07accuracy (%) | self-reported· optimizedT1 | 2023-06-27XGen-7B Technical Report |
Claims drawn from cited facts, not live model generation.
Developed by Salesforce AI Research and released in 2023, this model has 7 billion parameters. That places it at the compact end of the scale, where greater efficiency and lower cost come at the expense of the headroom larger models provide.
4 cited facts
This model is scored on six tracked benchmarks. It currently holds no top score on any of those benchmarks. The associated figures are self-reported and have not been independently confirmed, so they should be read as claims rather than verified results.
3 cited facts
Based on avg sota gap, it trails the leader by 51.62 points on average, a wide gap that places it well behind the leaders. It holds the top score on 0 benchmarks, meaning it has no category leadership on the tracked leaderboards.
2 cited facts
2 facts cross-checked across data sources: 2 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: curated_param_counts, epoch_benchmarks
XGen-7B is an AI model developed by Salesforce AI Research, released 2023-06-27. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
XGen-7B has recorded scores on 6 benchmarks, each shown with its evidence status.
XGen-7B has recorded scores on 6 benchmarks — PIQA, OpenBookQA, Winogrande, MMLU, ARC AI2, HellaSwag. The full table above shows each score with its evidence status.
0 of 6 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
XGen-7B has 6 tracked claims: 6 self-reported.