Add model-index evaluation metadata for LLaDA2.0-flash to README
#2
by
mackenzietechdocs
- opened
README.md
CHANGED
|
@@ -6,6 +6,101 @@ tags:
|
|
| 6 |
- diffusion
|
| 7 |
- llm
|
| 8 |
- text_generation
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
---
|
| 10 |
# LLaDA2.0-flash
|
| 11 |
|
|
|
|
| 6 |
- diffusion
|
| 7 |
- llm
|
| 8 |
- text_generation
|
| 9 |
+
model-index:
|
| 10 |
+
- name: LLaDA2.0-flash
|
| 11 |
+
results:
|
| 12 |
+
- task:
|
| 13 |
+
name: Text Generation
|
| 14 |
+
type: text-generation
|
| 15 |
+
dataset:
|
| 16 |
+
name: Benchmarks
|
| 17 |
+
type: benchmarks
|
| 18 |
+
metrics:
|
| 19 |
+
- name: Average
|
| 20 |
+
type: average
|
| 21 |
+
value: 79.32
|
| 22 |
+
|
| 23 |
+
# Knowledge
|
| 24 |
+
- name: MMLU
|
| 25 |
+
type: mmlu
|
| 26 |
+
value: 87.69
|
| 27 |
+
- name: MMLU-Pro
|
| 28 |
+
type: mmlu-pro
|
| 29 |
+
value: 73.36
|
| 30 |
+
- name: GPQA
|
| 31 |
+
type: gpqa
|
| 32 |
+
value: 61.98
|
| 33 |
+
- name: ARC-C
|
| 34 |
+
type: arc-c
|
| 35 |
+
value: 95.93
|
| 36 |
+
- name: CMMLU
|
| 37 |
+
type: cmmlu
|
| 38 |
+
value: 85.13
|
| 39 |
+
- name: C-EVAL
|
| 40 |
+
type: c-eval
|
| 41 |
+
value: 86.75
|
| 42 |
+
- name: GAOKAO-Bench
|
| 43 |
+
type: gaokao-bench
|
| 44 |
+
value: 93.90
|
| 45 |
+
|
| 46 |
+
# Reasoning
|
| 47 |
+
- name: SQuAD 2.0
|
| 48 |
+
type: squad-v2
|
| 49 |
+
value: 90.00
|
| 50 |
+
- name: DROP
|
| 51 |
+
type: drop
|
| 52 |
+
value: 87.90
|
| 53 |
+
- name: KOR-Bench
|
| 54 |
+
type: kor-bench
|
| 55 |
+
value: 64.24
|
| 56 |
+
- name: HellaSwag
|
| 57 |
+
type: hellaswag
|
| 58 |
+
value: 84.97
|
| 59 |
+
|
| 60 |
+
# Coding
|
| 61 |
+
- name: CRUXEval-O
|
| 62 |
+
type: cruxeval-o
|
| 63 |
+
value: 85.12
|
| 64 |
+
- name: MBPP
|
| 65 |
+
type: mbpp
|
| 66 |
+
value: 88.29
|
| 67 |
+
- name: MultiPL-E
|
| 68 |
+
type: multipl-e
|
| 69 |
+
value: 74.87
|
| 70 |
+
- name: HumanEval
|
| 71 |
+
type: humaneval
|
| 72 |
+
value: 94.51
|
| 73 |
+
- name: Bigcodebench-Full
|
| 74 |
+
type: bigcodebench-full
|
| 75 |
+
value: 41.58
|
| 76 |
+
- name: LiveCodeBench
|
| 77 |
+
type: livecodebench
|
| 78 |
+
value: 42.29
|
| 79 |
+
- name: Spider
|
| 80 |
+
type: spider
|
| 81 |
+
value: 82.49
|
| 82 |
+
|
| 83 |
+
# Math
|
| 84 |
+
- name: GSM8K
|
| 85 |
+
type: gsm8k
|
| 86 |
+
value: 96.06
|
| 87 |
+
- name: MATH
|
| 88 |
+
type: math
|
| 89 |
+
value: 95.44
|
| 90 |
+
- name: OlympiadBench
|
| 91 |
+
type: olympiadbench
|
| 92 |
+
value: 74.07
|
| 93 |
+
- name: AIME 2025
|
| 94 |
+
type: aime-2025
|
| 95 |
+
value: 60.00
|
| 96 |
+
|
| 97 |
+
# Agent & Alignment
|
| 98 |
+
- name: BFCL_Live
|
| 99 |
+
type: bfcl_live
|
| 100 |
+
value: 75.43
|
| 101 |
+
- name: IFEval-strict -prompt
|
| 102 |
+
type: ifeval-strict
|
| 103 |
+
value: 81.70
|
| 104 |
---
|
| 105 |
# LLaDA2.0-flash
|
| 106 |
|