From de2e89d21d6616439f94d7e59e69b811a751a5bc Mon Sep 17 00:00:00 2001 From: "Ramiro R. C." Date: Thu, 13 Nov 2025 07:48:20 -0300 Subject: [PATCH] Codeio prompt fix (#513) * prompr fix to request more specific JSON responses * corrected gallery examples too --- GALLERY.md | 6 +++--- reasoning_gym/code/codeio.py | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/GALLERY.md b/GALLERY.md index 6cfe9f26..b9bffd78 100644 --- a/GALLERY.md +++ b/GALLERY.md @@ -1668,7 +1668,7 @@ Given the following input: {'nums': [-94, 89, -30, -38]} -Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON object as your final answer. The keys and values of the object should strictly match the output requirement as specified. +Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON value (object, array, number or string, or one of the following three literal names: false null true) as your final answer. The keys and values of the object should strictly match the output requirement as specified. Tip: Here is a reference code snippet for this question. You can refer to this code to guide your reasoning but not copy spans of code directly. @@ -1706,7 +1706,7 @@ Given the following output: 0.001077807508543216 -Can you predict a feasible input without writing any code? Please reason and put your final answer in the form of a JSON object, even if the there is only one input variable, with keys strictly matching the input variables' names as specified. +Can you predict a feasible input without writing any code? Please reason and put your final answer in the form of a JSON value (object, array, number or string, or one of the following three literal names: false null true), even if the there is only one input variable, with keys strictly matching the input variables' names as specified. Tip: Here is a reference code snippet for this question. You can refer to this code to guide your reasoning but not copy spans of code directly. @@ -1798,7 +1798,7 @@ Given the following input: {'security_code': '923745', 'user_input': '623745'} -Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON object as your final answer. The keys and values of the object should strictly match the output requirement as specified. +Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON value (object, array, number or string, or one of the following three literal names: false null true) as your final answer. The keys and values of the object should strictly match the output requirement as specified. Tip: Here is a reference code snippet for this question. You can refer to this code to guide your reasoning but not copy spans of code directly. diff --git a/reasoning_gym/code/codeio.py b/reasoning_gym/code/codeio.py index 390b22a8..36b4469d 100644 --- a/reasoning_gym/code/codeio.py +++ b/reasoning_gym/code/codeio.py @@ -23,7 +23,7 @@ Given the following input: {2} -Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON object as your final answer. The keys and values of the object should strictly match the output requirement as specified. +Can you predict the output without writing any code? Please think and then provide the exact output in the form of a JSON value (object, array, number or string, or one of the following three literal names: false null true) as your final answer. The keys and values of the object should strictly match the output requirement as specified. Tip: Here is a reference code snippet for this question. You can refer to this code to guide your reasoning but not copy spans of code directly. @@ -43,7 +43,7 @@ Given the following output: {2} -Can you predict a feasible input without writing any code? Please reason and put your final answer in the form of a JSON object, even if the there is only one input variable, with keys strictly matching the input variables' names as specified. +Can you predict a feasible input without writing any code? Please reason and put your final answer in the form of a JSON value (object, array, number or string, or one of the following three literal names: false null true), even if the there is only one input variable, with keys strictly matching the input variables' names as specified. Tip: Here is a reference code snippet for this question. You can refer to this code to guide your reasoning but not copy spans of code directly.