HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

openchat-3.5-0106-GGUF

TheBloke/openchat-3.5-0106-GGUF

kind
gguf
quant
gguf (2-bit)
family
misc
modality
text-generation
root model
openchat/openchat-3.5-0106
provider
TheBloke
license
apache-2.0
language
—
context
8192,
library
transformers
vRAM (GB)
—
size on disk
—
downloads
1,076
likes
75
created
—
last modified
2024-01-10T16:24:15+00:00 (2024)
huggingface
https://huggingface.co/TheBloke/openchat-3.5-0106-GGUF
tags
transformersggufmistralopenchatC-RLFTtext-generationarxiv:2309.11235arxiv:2303.08774base_model:openchat/openchat-3.5-0106base_model:quantized:openchat/openchat-3.5-0106license:apache-2.0region:usconversational

Benchmark scores (11)

benchmarkcategoryscorecontextconfevidence
GSM8Kmath77.3BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
GSM8Kmath77.3high| OpenChat-3.5 | Apache-2.0 | **7B** | 56.4 | 64.3 | 55.5 | 28.6 | 77.3 |
MMLUknowledge64.3BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
MMLUknowledge64.3high| OpenChat-3.5 | Apache-2.0 | **7B** | 56.4 | 64.3 | 55.5 | 28.6 | 77.3 |
TruthfulQAsafety59.1BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
HumanEvalcoding55.5BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
HumanEvalcoding55.5high| OpenChat-3.5 | Apache-2.0 | **7B** | 56.4 | 64.3 | 55.5 | 28.6 | 77.3 |
BBHreasoning47.6BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
AGIEvalreasoning47.4BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |
MATHmath28.6high| OpenChat-3.5 | Apache-2.0 | **7B** | 56.4 | 64.3 | 55.5 | 28.6 | 77.3 |
MT-Benchchat7.81BBH CoThigh| OpenChat-3.5 | **7B** | 61.6 | 7.81 | 55.5 | 47.6 | 47.4 | 59.1 | 64.3 | 77.3 | 63.5 |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.