bytedance / bytedance/FullStackBench

Confusion about evaluation datasets

Open
#3 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
121
Forks
10
PR merge metrics
No merged PRs in 30d

Description

When I was looking at the evaluation questions, I found that some PHP questions required the code to be converted into PHP, but the reference code provided was already PHP. Is this intentionally designed to test the model copying ability or is it a design error?

Two examples are as follows:
```
"content": "Please convert the following code to PHP, with the function signature `function order_by_points($nums)`\nNote:\n1. Only implement the target function and import necessary packages, do not generate other code\n\n```php\n\nfunction order_by_points($nums) {\n \n function digits_sum($n) {\n $neg = 1;\n if ($n < 0) {\n $n = -1 * $n;\n $neg = -1;\n }\n $n = array_map('intval', str_split($n));\n $n[0] = $n[0] * $neg;\n return array_sum($n);\n }\n usort($nums, function($a, $b) {\n return digits_sum($a) <=> digits_sum($b);\n });\n return $nums;\n}\n\n```",
"id": 297,

"content": "Please convert the following code to PHP, with the function signature `function do_algebra($operator, $operand)`. Note: Only implement the target function and import necessary packages, do not generate other code.\n\n```php\n\nfunction do_algebra($operator, $operand) {\n\n $expression = strval($operand[0]);\n foreach (array_combine($operator, array_slice($operand, 1)) as $oprt => $oprn) {\n $expression .= $oprt . strval($oprn);\n }\n return eval('return '.$expression.';');\n}\n\n```",
"id": 298,
```

Contributor guide

No contributing guide indexed for this repository

Research direction

Inspect evaluation dataset entries 297 and 298, which are the two examples named in the issue, and compare their PHP conversion prompts with the embedded reference code. Done means establishing whether the prompt/reference mismatch is intentional or correcting the affected dataset entries accordingly.

Written by the indexing model from the issue text.

Assessment

Tech stack
php
Domain
testing-qa
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.