Hugging Face's logo

euclaise
/
Memphis-scribe-3B

Text Generation
Transformers
Safetensors
English
stablelm_epoch
supertrainer2000
Not-For-All-Audiences
writing
roleplay
conversational
custom_code
Model card Files Community
euclaise commited on
Commit
7ee2243
·
verified ·
1 Parent(s): 054b3db

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -59,6 +59,6 @@ This model performs significantly worse than Memphis-CoT on benchmarks, despite
59
  | Model | GSM8K (5-shot) | AGIEval (English/Nous subset, acc_norm) | BIG Bench Hard (CoT, few-shot*) |
60
  |:---------------------------------------------------------------------------|:---------------|:----------------------------------------|:--------------------------------|
61
  | [StableLM 3B Base](https://hf.co/stabilityai/stablelm-3b-4e1t) | 2.05% | 25.14% | 36.75% |
62
- | [Memphis-CoT 3B](https://hf.co/euclaise/Memphis-CoT-3B) | 8.8% | 27.22% | 36.92% |
63
  | [Memphis-scribe 3B](https://hf.co/euclaise/Memphis-scribe-3B) | 9.55% | 24.78% | |
64
  *5-shot, as performed automatically by LM Evaluation Harness bbh_cot_fewshot even with num_fewshot=0
 
59
  | Model | GSM8K (5-shot) | AGIEval (English/Nous subset, acc_norm) | BIG Bench Hard (CoT, few-shot*) |
60
  |:---------------------------------------------------------------------------|:---------------|:----------------------------------------|:--------------------------------|
61
  | [StableLM 3B Base](https://hf.co/stabilityai/stablelm-3b-4e1t) | 2.05% | 25.14% | 36.75% |
62
+ | [Memphis-CoT 3B](https://hf.co/euclaise/Memphis-CoT-3B) | 18.8% | 27.22% | 36.92% |
63
  | [Memphis-scribe 3B](https://hf.co/euclaise/Memphis-scribe-3B) | 9.55% | 24.78% | |
64
  *5-shot, as performed automatically by LM Evaluation Harness bbh_cot_fewshot even with num_fewshot=0