Datasets:

zeyu12
/

eval

Modalities:

Text

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

test · 200 rows

idx int64 0 99	question stringlengths 21 1.49k	gt_cot stringlengths 55 1.49k	gt stringlengths 1 15	level stringclasses 5 values	solution stringlengths 55 1.49k	answer stringlengths 1 23	code listlengths 8 8	pred listlengths 8 8	report listlengths 8 8	score listlengths 8 8	system stringclasses 1 value	query stringlengths 21 1.49k	response listlengths 8 8	history listlengths 0 0	pred_score listlengths 8 8
0	What is $10.0000198\cdot 5.9999985401\cdot 6.9999852$ to the nearest whole number?	Notice that $10.00001988$ is very close to $10$, $5.9999985401$ is very close to $6$ and $6.9999852$ is very close to $7$. Because the given numbers are all so close to integers, we're unlikely to go wrong by rounding before multiplying. We get $$10\cdot6\cdot7=\boxed{420}.$$If we multiplied the given numbers with a ca...	420	Level 3	Notice that $10.00001988$ is very close to $10$, $5.9999985401$ is very close to $6$ and $6.9999852$ is very close to $7$. Because the given numbers are all so close to integers, we're unlikely to go wrong by rounding before multiplying. We get $$10\cdot6\cdot7=\boxed{420}.$$If we multiplied the given numbers with a ca...	420	[ "To find the value of \$10.0000198 \\cdot 5.9999985401 \\cdot 6.9999852\$ to the nearest whole number, we can approximate each number to a simpler form that is easier to work with. Since each of these numbers is very close to a whole number, we can approximate them as follows:\n\n- \$10.0000198 \\approx 10\$\n-...	[ "420", "420", "420", "420", "420", "420", "420", "420" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	What is $10.0000198\cdot 5.9999985401\cdot 6.9999852$ to the nearest whole number?	[ "To find the value of \$10.0000198 \\cdot 5.9999985401 \\cdot 6.9999852\$ to the nearest whole number, we can approximate each number to a simpler form that is easier to work with. Since each of these numbers is very close to a whole number, we can approximate them as follows:\n\n- \$10.0000198 \\approx 10\$\n-...	[]	[ [ 3.546875 ], [ 3.515625 ], [ 3.46875 ], [ 3.46875 ], [ 3.265625 ], [ 3.328125 ], [ 3.5 ], [ 3.4375 ] ]
1	Two fair 6-sided dice are rolled. What is the probability the sum rolled is 9?	There are 4 different ways to roll a 9 (3+6, 4+5, 5+4, 6+3), which makes the probability of rolling a 9 equal to $\dfrac{4}{36} = \boxed{\dfrac{1}{9}}$.	\frac{1}{9}	Level 4	There are 4 different ways to roll a 9 (3+6, 4+5, 5+4, 6+3), which makes the probability of rolling a 9 equal to $\dfrac{4}{36} = \boxed{\dfrac{1}{9}}$.	\dfrac{1}{9}	[ "To determine the probability that the sum of the numbers rolled on two fair 6-sided dice is 9, we need to follow these steps:\n\n1. Identify all possible outcomes when two dice are rolled:\n Each die has 6 faces, and the outcome of the first die is independent of the outcome of the second die. Therefore, the...	[ "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}", "\\frac{1}{9}" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	Two fair 6-sided dice are rolled. What is the probability the sum rolled is 9?	[ "To determine the probability that the sum of the numbers rolled on two fair 6-sided dice is 9, we need to follow these steps:\n\n1. Identify all possible outcomes when two dice are rolled:\n Each die has 6 faces, and the outcome of the first die is independent of the outcome of the second die. Therefore, the...	[]	[ [ 5.375 ], [ 5.375 ], [ 5.5 ], [ 5.46875 ], [ 5.4375 ], [ 5.4375 ], [ 5.5 ], [ 5.5625 ] ]
2	For what value of $x$ is the following equation true: $6500 + x - 4500 = 3400 + 2000$?	Simplifying both sides gives $x +2000 = 5400$. Subtracting 2000 from both sides gives $x = \boxed{3400}$.	3400	Level 2	Simplifying both sides gives $x +2000 = 5400$. Subtracting 2000 from both sides gives $x = \boxed{3400}$.	3400	[ "To solve the equation \$6500 + x - 4500 = 3400 + 2000\$, we will follow these steps:\n\n1. Simplify both sides of the equation by combining like terms.\n2. Isolate the variable \$x\$ on one side of the equation.\n\nLet's start with the left side of the equation:\n\\[6500 + x - 4500\\]\nCombine the constants:\n...	[ "3400", "3400", "3400", "3400", "3400", "3400", "3400", "3400" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	For what value of $x$ is the following equation true: $6500 + x - 4500 = 3400 + 2000$?	[ "To solve the equation \$6500 + x - 4500 = 3400 + 2000\$, we will follow these steps:\n\n1. Simplify both sides of the equation by combining like terms.\n2. Isolate the variable \$x\$ on one side of the equation.\n\nLet's start with the left side of the equation:\n\\[6500 + x - 4500\\]\nCombine the constants:\n...	[]	[ [ 4.59375 ], [ 4.53125 ], [ 4.59375 ], [ 4.59375 ], [ 4.625 ], [ 4.53125 ], [ 4.4375 ], [ 4.5625 ] ]
3	At 2:30 p.m. during a long drive, Bobbi asks her parents, ``Are we there yet?'' Her mother responds, ``We will be there in 7200 seconds.'' If Bobbi's mother is correct, at what time in the afternoon will they arrive at their destination?	Multiply 7200 seconds by $\left(\frac{1\text{ min.}}{60\text{ sec.}}\right)\left(\frac{1\text{ hr.}}{60\text{ min.}}\right)$ to find that they will arrive in 2 hours. Two hours after 2:30 p.m. is $\boxed{\text{4:30 p.m.}}$.	4:30p..	Level 3	Multiply 7200 seconds by $\left(\frac{1\text{ min.}}{60\text{ sec.}}\right)\left(\frac{1\text{ hr.}}{60\text{ min.}}\right)$ to find that they will arrive in 2 hours. Two hours after 2:30 p.m. is $\boxed{\text{4:30 p.m.}}$.	\text{4:30 p.m.}	[ "To determine the time at which Bobbi and her family will arrive at their destination, we need to add the travel time to the current time. The current time is 2:30 p.m., and the travel time is 7200 seconds. We will convert the travel time from seconds to hours, and then add this time to the current time.\n\nFirst, ...	[ "4:30", "4:30", "4:30", "4:30", "4:30", "4:30", "4:30", "4:30" ]	[ null, null, null, null, null, null, null, null ]	[ false, false, false, false, false, false, false, false ]	Please reason step by step, and put your final answer within \boxed{}.	At 2:30 p.m. during a long drive, Bobbi asks her parents, ``Are we there yet?'' Her mother responds, ``We will be there in 7200 seconds.'' If Bobbi's mother is correct, at what time in the afternoon will they arrive at their destination?	[ "To determine the time at which Bobbi and her family will arrive at their destination, we need to add the travel time to the current time. The current time is 2:30 p.m., and the travel time is 7200 seconds. We will convert the travel time from seconds to hours, and then add this time to the current time.\n\nFirst, ...	[]	[ [ 0.62109375 ], [ 1.4921875 ], [ 1.4609375 ], [ 1.4375 ], [ 1.203125 ], [ 0.46875000000000006 ], [ 1.234375 ], [ 1.390625 ] ]
4	Compute $ \frac{1}{31} \div \frac{1}{93} $.	Recall that division is multiplication by a reciprocal. So we rewrite the division by $\frac{1}{93}$ as multiplication by the reciprocal of $\frac{1}{93}$. Since the reciprocal of $\frac{1}{93}$ is 93, we get \[ \frac{1}{31} \div \frac{1}{93} = \frac{1}{31} \cdot 93. \]Now we can use the commutative property of multip...	3	Level 1	Recall that division is multiplication by a reciprocal. So we rewrite the division by $\frac{1}{93}$ as multiplication by the reciprocal of $\frac{1}{93}$. Since the reciprocal of $\frac{1}{93}$ is 93, we get \[ \frac{1}{31} \div \frac{1}{93} = \frac{1}{31} \cdot 93. \]Now we can use the commutative property of multip...	3	[ "To compute \$ \\frac{1}{31} \\div \\frac{1}{93} \$, we start by understanding that dividing by a fraction is equivalent to multiplying by its reciprocal. The reciprocal of \$ \\frac{1}{93} \$ is \$ \\frac{93}{1} \$. Therefore, we can rewrite the division as a multiplication:\n\n\\[\n\\frac{1}{31} \\div \\fra...	[ "3", "3", "3", "3", "3", "3", "3", "3" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	Compute $ \frac{1}{31} \div \frac{1}{93} $.	[ "To compute \$ \\frac{1}{31} \\div \\frac{1}{93} \$, we start by understanding that dividing by a fraction is equivalent to multiplying by its reciprocal. The reciprocal of \$ \\frac{1}{93} \$ is \$ \\frac{93}{1} \$. Therefore, we can rewrite the division as a multiplication:\n\n\\[\n\\frac{1}{31} \\div \\fra...	[]	[ [ 3.515625 ], [ 3.71875 ], [ 3.765625 ], [ 3.84375 ], [ 3.75 ], [ 3.828125 ], [ 3.734375 ], [ 3.78125 ] ]
5	What is the area, in square units, of an isosceles right triangle with a hypotenuse of 20 units?	Each leg of a 45-45-90 triangle with a hypotenuse of 20 units measures $\frac{20}{\sqrt{2}}$ units. The area is $\frac{1}{2}(\text{base})(\text{height})=\frac{1}{2}\left(\frac{20}{\sqrt{2}}\right)\left(\frac{20}{\sqrt{2}}\right)=\frac{400}{2\cdot 2}=\boxed{100\text{ square units}}$.	100	Level 5	Each leg of a 45-45-90 triangle with a hypotenuse of 20 units measures $\frac{20}{\sqrt{2}}$ units. The area is $\frac{1}{2}(\text{base})(\text{height})=\frac{1}{2}\left(\frac{20}{\sqrt{2}}\right)\left(\frac{20}{\sqrt{2}}\right)=\frac{400}{2\cdot 2}=\boxed{100\text{ square units}}$.	100\text{ square units}	[ "To find the area of an isosceles right triangle with a hypotenuse of 20 units, we start by recalling the properties of an isosceles right triangle. In such a triangle, the two legs are of equal length, and the relationship between the legs and the hypotenuse is given by the Pythagorean theorem. Specifically, if th...	[ "100", "100", "100", "100", "100", "100", "100", "100" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	What is the area, in square units, of an isosceles right triangle with a hypotenuse of 20 units?	[ "To find the area of an isosceles right triangle with a hypotenuse of 20 units, we start by recalling the properties of an isosceles right triangle. In such a triangle, the two legs are of equal length, and the relationship between the legs and the hypotenuse is given by the Pythagorean theorem. Specifically, if th...	[]	[ [ 1.671875 ], [ 1.5078125 ], [ 1.3828125 ], [ 1.546875 ], [ 1.8359375 ], [ 1.5625 ], [ 1.609375 ], [ 1.5703125 ] ]
6	The square quilt block shown is used in a larger quilt. The four small squares in the block are congruent, and the four small vertical rectangles in the block that are not squares are also congruent. The total area of the shaded regions is what fraction of the total area of the block? Express your answer as a common fr...	Suppose the whole block has area $1$. Each of the smaller squares then has area $1/4$. Each big right triangle has area $(1/2)(1/4)=1/8$, while each small right triangle has area $(1/8)(1/4)=1/32$. Thus the total shaded area, which equals the fraction covered, equals $4(1/32)+2(1/8)=\boxed{\frac{3}{8}}$.	\frac{3}{8}	Level 4	Suppose the whole block has area $1$. Each of the smaller squares then has area $1/4$. Each big right triangle has area $(1/2)(1/4)=1/8$, while each small right triangle has area $(1/8)(1/4)=1/32$. Thus the total shaded area, which equals the fraction covered, equals $4(1/32)+2(1/8)=\boxed{\frac{3}{8}}$.	\frac{3}{8}	[ "To determine the fraction of the total area of the quilt block that is shaded, we start by analyzing the structure of the block. The quilt block is a square divided into smaller squares and rectangles. Let's denote the side length of the entire quilt block as \$ s \$.\n\nSince the block is divided into four smal...	[ "\\frac{5}{16}", "\\frac{3}{8}", "\\frac{3}{8}", "\\frac{3}{4}", "\\frac{3}{8}", "\\frac{1}{4}", "\\frac{3}{16}", "\\frac{5}{16}" ]	[ null, null, null, null, null, null, null, null ]	[ false, true, true, false, true, false, false, false ]	Please reason step by step, and put your final answer within \boxed{}.	The square quilt block shown is used in a larger quilt. The four small squares in the block are congruent, and the four small vertical rectangles in the block that are not squares are also congruent. The total area of the shaded regions is what fraction of the total area of the block? Express your answer as a common fr...	[ "To determine the fraction of the total area of the quilt block that is shaded, we start by analyzing the structure of the block. The quilt block is a square divided into smaller squares and rectangles. Let's denote the side length of the entire quilt block as \$ s \$.\n\nSince the block is divided into four smal...	[]	[ [ -2.6875 ], [ 0.056396484375 ], [ 0.36328125 ], [ -0.52734375 ], [ 0.162109375 ], [ -1.21875 ], [ -1.7890625 ], [ -2.28125 ] ]
7	Robin bought a four-scoop ice cream cone having a scoop each of vanilla, chocolate, strawberry and cherry. In how many orders can the four scoops be stacked on the cone if they are stacked one on top of the other and chocolate cannot be on the bottom?	There are 3 choices for the bottom scoop. Then there are only 3 choices for the scoop above that (since one flavor has already been used), then 2 choices for the next scoop, and 1 choice for the final scoop. This gives a total of $3\cdot 3\cdot 2\cdot 1 = \boxed{18}$ possible cones.	18	Level 4	There are 3 choices for the bottom scoop. Then there are only 3 choices for the scoop above that (since one flavor has already been used), then 2 choices for the next scoop, and 1 choice for the final scoop. This gives a total of $3\cdot 3\cdot 2\cdot 1 = \boxed{18}$ possible cones.	18	[ "To determine the number of orders in which the four scoops of ice cream can be stacked on the cone, with the condition that the chocolate scoop cannot be on the bottom, we can follow these steps:\n\n1. Calculate the total number of permutations of the four scoops without any restrictions.\n Since there are f...	[ "18", "18", "18", "18", "18", "18", "18", "18" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	Robin bought a four-scoop ice cream cone having a scoop each of vanilla, chocolate, strawberry and cherry. In how many orders can the four scoops be stacked on the cone if they are stacked one on top of the other and chocolate cannot be on the bottom?	[ "To determine the number of orders in which the four scoops of ice cream can be stacked on the cone, with the condition that the chocolate scoop cannot be on the bottom, we can follow these steps:\n\n1. Calculate the total number of permutations of the four scoops without any restrictions.\n Since there are f...	[]	[ [ 3.265625 ], [ 3.15625 ], [ 3.40625 ], [ 3.1875 ], [ 3.203125 ], [ 3.390625 ], [ 3.328125 ], [ 3.203125 ] ]
8	What is the positive difference between the greatest and the least member of the set $\left\{\frac{3}{7},\frac{4}{3},\frac{11}{8},\frac{6}{16}\right\}$? Express your answer in simplest form.	The least member of the set is either $\frac{3}{7}$ or $\frac{6}{16}$, since these two members are the only ones smaller than 1. Since $\frac{3}{7}=\frac{6}{14}$, $\frac{6}{16}$ is the smaller of the two. The greatest member is either $\frac{4}{3}$ or $\frac{11}{8}$. Since $\frac{4}{3}=\frac{32}{24}$ and $\frac{11}{8}=...	1	Level 4	The least member of the set is either $\frac{3}{7}$ or $\frac{6}{16}$, since these two members are the only ones smaller than 1. Since $\frac{3}{7}=\frac{6}{14}$, $\frac{6}{16}$ is the smaller of the two. The greatest member is either $\frac{4}{3}$ or $\frac{11}{8}$. Since $\frac{4}{3}=\frac{32}{24}$ and $\frac{11}{8}=...	1	[ "To find the positive difference between the greatest and the least member of the set \$\\left\\{\\frac{3}{7}, \\frac{4}{3}, \\frac{11}{8}, \\frac{6}{16}\\right\\}\$, we first need to determine the value of each fraction and then identify the greatest and the least of these values.\n\nFirst, let's simplify the fr...	[ "1", "1", "1", "1", "1", "1", "1", "1" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	What is the positive difference between the greatest and the least member of the set $\left\{\frac{3}{7},\frac{4}{3},\frac{11}{8},\frac{6}{16}\right\}$? Express your answer in simplest form.	[ "To find the positive difference between the greatest and the least member of the set \$\\left\\{\\frac{3}{7}, \\frac{4}{3}, \\frac{11}{8}, \\frac{6}{16}\\right\\}\$, we first need to determine the value of each fraction and then identify the greatest and the least of these values.\n\nFirst, let's simplify the fr...	[]	[ [ 3.390625 ], [ 3.171875 ], [ 3.109375 ], [ 3.390625 ], [ 3.234375 ], [ 3.046875 ], [ 3.390625 ], [ 3.359375 ] ]
9	If a certain negative number is multiplied by six, the result is the same as 20 less than the original number. What is the value of the original number?	Let the number be $x$. We know that $6x=x-20$. Subtracting $x$ from both sides gives $5x=-20$. Then, dividing both sides by 5 gives $x= \boxed{-4}$.	-4	Level 3	Let the number be $x$. We know that $6x=x-20$. Subtracting $x$ from both sides gives $5x=-20$. Then, dividing both sides by 5 gives $x= \boxed{-4}$.	-4	[ "To solve the problem, let's denote the original negative number by \$ x \$. According to the problem, when this number is multiplied by 6, the result is the same as 20 less than the original number. We can express this relationship with the following equation:\n\n\\[ 6x = x - 20 \\]\n\nNext, we need to isolate \...	[ "-4", "-4", "-4", "-4", "-4", "-4", "-4", "-4" ]	[ null, null, null, null, null, null, null, null ]	[ true, true, true, true, true, true, true, true ]	Please reason step by step, and put your final answer within \boxed{}.	If a certain negative number is multiplied by six, the result is the same as 20 less than the original number. What is the value of the original number?	[ "To solve the problem, let's denote the original negative number by \$ x \$. According to the problem, when this number is multiplied by 6, the result is the same as 20 less than the original number. We can express this relationship with the following equation:\n\n\\[ 6x = x - 20 \\]\n\nNext, we need to isolate \...	[]	[ [ 4.875 ], [ 4.90625 ], [ 4.90625 ], [ 4.90625 ], [ 4.875 ], [ 4.9375 ], [ 4.9375 ], [ 4.90625 ] ]

End of preview. Expand in Data Studio

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

eval — 数学 / 代码 / 通用能力评测自包含工具包

从集群各框架抽取的可独立运行评测套件（代码 + 数据全齐），三个子目录互不依赖。

eval/
├── math/                      # 数学评测 (vllm_eval.py, 生成式, greedy)
│   ├── run_math.sh            #   入口: sbatch 提交 (MODEL=... )
│   ├── vllm_eval.py           #   主评测器 (vLLM 后端)
│   ├── data_loader.py / evaluate.py / parser.py / utils.py / ...   # 判分与数据链路
│   ├── latex2sympy/           #   本地判分包 (首次运行自动 pip install -e)
│   └── data/                  #   数据 (含默认 7 基准 + 上游全套 24 个数据集)
├── general/                   # 通用能力评测 (general_eval.py, 官方协议)
│   ├── run_general.sh         #   入口: sbatch 提交 (MODEL=..., CHAT=1 默认)
│   ├── general_eval.py        #   主评测器 (vLLM 后端, 无本地依赖)
│   ├── download_data.py       #   数据重建脚本 (数据已随包附带)
│   └── data/                  #   hellaswag / arc_challenge / mmlu_pro / tydiqa / bbh
└── code/                      # 代码评测 (EvalPlus: HumanEval+ / MBPP+)
    ├── run_evalplus.sh        #   入口: sbatch 提交 (MODEL=... ) 或 bash 传参
    ├── summarize_results.py   #   批量结果汇总工具
    ├── run_evalplus.sh.orig   #   原始脚本参考 (绝对路径版)
    └── data/                  #   HumanEvalPlus-v0.1.10.jsonl / MbppPlus-v0.2.0.jsonl
                                #   (运行时自动种子到 ~/.cache/evalplus, 离线可用)

依赖（集群上已就绪）

conda 环境 vllm：/data/icto/mc56499/miniconda3/envs/vllm（含 vllm、evalplus.*、判分依赖）
模块：module load nvidia cuda/13.0.2（脚本内已自带，flashinfer JIT 需要 nvcc）
三个入口脚本都自带 unset LD_LIBRARY_PATH LD_PRELOAD 环境防护（IDE 终端提交免疫）

用法

数学（7 基准：gsm8k/math500/aime24/aime25/amc2023/minerva_math/olympiadbench）

MODEL=/data/icto/mc56499/model/Qwen2.5-7B \
    sbatch /data/icto/mc56499/eval/math/run_math.sh          # pro6000 2卡

MODEL=... sbatch -p gpu-l40s /data/icto/mc56499/eval/math/run_math.sh   # 换分区
DATA_NAMES=gsm8k,math500 MODEL=... sbatch ...                 # 只跑子集

口径与 math_evaluation 完全一致：--prompt_type qwen-boxed、temperature 0、 n_sampling 1、max_tokens 3072、max_model_len 4096、TP 自动 = 可见 GPU 数。结果：math/results/<模型名>/<时间戳>/summary.json；已有 summary 自动跳过（断点续跑）。

通用能力（5 基准：HellaSwag / ARC-Challenge / MMLU-Pro / TyDiQA-GoldP / BBH）

MODEL=/data/icto/mc56499/model/Qwen2.5-7B \
    sbatch /data/icto/mc56499/eval/general/run_general.sh    # pro6000 2卡, chat 模板默认开

CHAT=0 MODEL=... sbatch ...                                   # base 模型关模板
DATA_NAMES=mmlu_pro MODEL=... sbatch ...                      # 只跑子集

口径与 general_evaluation 完全一致：官方协议、greedy、VLLM_USE_FLASHINFER_SAMPLER=0、 BBH 官方 3-shot CoT。结果：general/results/<模型名>/<时间戳>/summary.json，断点续跑同上。

代码（HumanEval+ / MBPP+，pass@1 base & plus）

sbatch --export=ALL,MODEL=/data/icto/mc56499/model/Qwen2.5-Coder-3B \
    /data/icto/mc56499/eval/code/run_evalplus.sh              # pro6000 2卡

# 或交互式: bash run_evalplus.sh /path/to/model [TP]

结果：code/evalplus_results/<dataset>/（样本 jsonl + *_eval_results.json），脚本末尾自动打印 pass@1 (base tests) / pass@1 (base+extra tests)。

数据来源

类别	数据集	版本/说明
数学	gsm8k, math500, aime24, aime25, amc2023, minerva_math, olympiadbench	默认 7 基准，`math/data/` 下另有 aqua/asdiv/cmath/gaokao 等 24 个上游数据集，改 `DATA_NAMES` 即可用
通用	HellaSwag, ARC-Challenge, MMLU-Pro, TyDiQA-GoldP, BBH	官方协议；BBH 含 27 个官方 3-shot CoT prompt
代码	HumanEval+	v0.1.10 (164 题)
代码	MBPP+	v0.2.0 (378 题)

注意

评测产物（results/、evalplus_results/、logs/）写在本目录内，不污染外部框架
math/README_upstream.md 为上游框架原始说明，供参考
换机器运行：需自备 vllm + evalplus + latex2sympy 环境（pip install -e math/latex2sympy）

Downloads last month: 75

eval — 数学 / 代码 / 通用能力 评测自包含工具包