Hugging Face's logo

RLHFlow
/
LLaMA3-iterative-DPO-final

Text Generation
Transformers
Safetensors
llama
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
3

Adding Evaluation Results

#3
by leaderboard-pr-bot - opened
base: refs/heads/main
←
from: refs/pr/3
Discussion Files changed
+109
-1
Files changed (1) hide show
  1. README.md +109 -1
README.md CHANGED
@@ -1,5 +1,100 @@
1
  ---
2
  license: llama3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
4
  # LLaMA3-iterative-DPO-final
5
 
@@ -117,4 +212,17 @@ Please cite our techical report if you find our model is useful for your researc
117
  primaryClass={cs.LG}
118
  }
119
 
120
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: llama3
3
+ model-index:
4
+ - name: LLaMA3-iterative-DPO-final
5
+ results:
6
+ - task:
7
+ type: text-generation
8
+ name: Text Generation
9
+ dataset:
10
+ name: IFEval (0-Shot)
11
+ type: HuggingFaceH4/ifeval
12
+ args:
13
+ num_few_shot: 0
14
+ metrics:
15
+ - type: inst_level_strict_acc and prompt_level_strict_acc
16
+ value: 53.34
17
+ name: strict accuracy
18
+ source:
19
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
20
+ name: Open LLM Leaderboard
21
+ - task:
22
+ type: text-generation
23
+ name: Text Generation
24
+ dataset:
25
+ name: BBH (3-Shot)
26
+ type: BBH
27
+ args:
28
+ num_few_shot: 3
29
+ metrics:
30
+ - type: acc_norm
31
+ value: 29.79
32
+ name: normalized accuracy
33
+ source:
34
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
35
+ name: Open LLM Leaderboard
36
+ - task:
37
+ type: text-generation
38
+ name: Text Generation
39
+ dataset:
40
+ name: MATH Lvl 5 (4-Shot)
41
+ type: hendrycks/competition_math
42
+ args:
43
+ num_few_shot: 4
44
+ metrics:
45
+ - type: exact_match
46
+ value: 0.0
47
+ name: exact match
48
+ source:
49
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
50
+ name: Open LLM Leaderboard
51
+ - task:
52
+ type: text-generation
53
+ name: Text Generation
54
+ dataset:
55
+ name: GPQA (0-shot)
56
+ type: Idavidrein/gpqa
57
+ args:
58
+ num_few_shot: 0
59
+ metrics:
60
+ - type: acc_norm
61
+ value: 4.47
62
+ name: acc_norm
63
+ source:
64
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
65
+ name: Open LLM Leaderboard
66
+ - task:
67
+ type: text-generation
68
+ name: Text Generation
69
+ dataset:
70
+ name: MuSR (0-shot)
71
+ type: TAUR-Lab/MuSR
72
+ args:
73
+ num_few_shot: 0
74
+ metrics:
75
+ - type: acc_norm
76
+ value: 5.08
77
+ name: acc_norm
78
+ source:
79
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
80
+ name: Open LLM Leaderboard
81
+ - task:
82
+ type: text-generation
83
+ name: Text Generation
84
+ dataset:
85
+ name: MMLU-PRO (5-shot)
86
+ type: TIGER-Lab/MMLU-Pro
87
+ config: main
88
+ split: test
89
+ args:
90
+ num_few_shot: 5
91
+ metrics:
92
+ - type: acc
93
+ value: 25.08
94
+ name: accuracy
95
+ source:
96
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=RLHFlow/LLaMA3-iterative-DPO-final
97
+ name: Open LLM Leaderboard
98
  ---
99
  # LLaMA3-iterative-DPO-final
100
 
 
212
  primaryClass={cs.LG}
213
  }
214
 
215
+ ```
216
+ # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard)
217
+ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_RLHFlow__LLaMA3-iterative-DPO-final)
218
+
219
+ | Metric |Value|
220
+ |-------------------|----:|
221
+ |Avg. |19.96|
222
+ |IFEval (0-Shot) |53.34|
223
+ |BBH (3-Shot) |29.79|
224
+ |MATH Lvl 5 (4-Shot)| 0.00|
225
+ |GPQA (0-shot) | 4.47|
226
+ |MuSR (0-shot) | 5.08|
227
+ |MMLU-PRO (5-shot) |25.08|
228
+