Hugging Face's logo

Datasets:
Dhevenddra
/
slm-lab-data

Modalities:
Text
Formats:
json
Languages:
English
Size:
10K - 100K
Tags:
educational
tokenizers
Synthetic
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Dhevenddra commited on
Commit
2b5cf15
·
verified ·
1 Parent(s): df258b8

add phase09_benchmarks.md

Browse files
Files changed (1) hide show
  1. results/phase09_benchmarks.md +1 -1
results/phase09_benchmarks.md CHANGED
@@ -5,7 +5,7 @@
5
  | model | params | hellaswag | arc_easy | arc_challenge | piqa | mean above chance |
6
  |---|---|---|---|---|---|---|
7
  | Phase 03 base (12.6M, from scratch) | 12,587,904 | 0.2619 | 0.2572 | 0.2201 | 0.5196 | **+0.0022** |
8
- | Phase 05 SFT (12.6M, from scratch) | 12,587,904 | not run | not run | not run | not run | not run |
9
  | Llama-3.2-1B-Instruct (base) | 1,235,814,400 | not run | not run | not run | not run | not run |
10
  | Phase 06 QLoRA fine-tune (r=8) | 1,235,814,400 | not run | not run | not run | not run | not run |
11
  | Phase 07 DPO (beta=0.1) | 1,235,814,400 | not run | not run | not run | not run | not run |
 
5
  | model | params | hellaswag | arc_easy | arc_challenge | piqa | mean above chance |
6
  |---|---|---|---|---|---|---|
7
  | Phase 03 base (12.6M, from scratch) | 12,587,904 | 0.2619 | 0.2572 | 0.2201 | 0.5196 | **+0.0022** |
8
+ | Phase 05 SFT (12.6M, from scratch) | 12,587,904 | 0.2621 | 0.2601 | 0.2346 | 0.5180 | **+0.0062** |
9
  | Llama-3.2-1B-Instruct (base) | 1,235,814,400 | not run | not run | not run | not run | not run |
10
  | Phase 06 QLoRA fine-tune (r=8) | 1,235,814,400 | not run | not run | not run | not run | not run |
11
  | Phase 07 DPO (beta=0.1) | 1,235,814,400 | not run | not run | not run | not run | not run |