Dataset Viewer
Auto-converted to Parquet Duplicate
sample_id
string
parsed_answer
string
correct_answer
null
stop_reason
string
output_tokens
int64
input_tokens
int64
reasoning
string
prompt
list
ex00007
0.999
null
stop
1,070
366
Let me analyze this code carefully. **Specification:** - Function name: `second_largest` - Input: `nums` (list of numbers) - Output: the second largest distinct value in `nums` - Exception: `ValueError` if all elements are equal (no distinct second largest exists) **Implementation:** ```python def second_largest(nums...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00024
0.001
null
stop
752
323
Let me analyze this code carefully. The specification says: - Implement `median3(a, b, c)` returning the median of three distinct integers. The implementation is: ```python def median3(a, b, c): ordered = tuple(sorted((a, b, c))) return ordered[2] ``` Let's think about what the median is. Given three numbers...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00029
0.999
null
stop
805
387
Let me analyze this code carefully. The specification says: - Return a sorted list of unique values from `nums` The implementation: ```python def unique_sorted(nums): remaining = set(nums) result = [] while remaining: smallest = min(remaining) result.append(smallest) remaining.remo...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00043
0.001
null
stop
842
399
Let me analyze this code carefully. **Specification:** - Find the first element strictly greater than `threshold` - Return `None` if no such element exists **Implementation:** ```python def find_first_gt(nums, threshold): matches = [n for n in nums if n >= threshold] return matches[0] if matches else None ```...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00044
0.001
null
stop
1,078
405
Let me carefully analyze this code to determine if it correctly implements the clamp function. The specification says: - `clamp(x, lo, hi)` should return `x` clamped into the closed interval `[lo, hi]` - Assume `lo <= hi` So: - If `x < lo`, return `lo` - If `x > hi`, return `hi` - Otherwise (if `lo <= x <= hi`), retu...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00050
0.999
null
stop
995
381
Let me analyze this code carefully. The specification asks for: - `unique_sorted(nums)` should return a sorted list of unique values from `nums` Let's trace through the implementation: ```python def unique_sorted(nums): remaining = set(nums) result = [] while remaining: smallest = min(remaining) ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00052
0.999
null
stop
1,897
406
Let me analyze this code carefully. The specification requires: 1. Return the value with strictly highest count 2. If the top count is tied, raise ValueError Let's trace through the implementation: ```python def majority_vote(labels): if not labels: raise ValueError('empty') from collections import C...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00059
0.999
null
stop
909
342
Let me carefully analyze this code. The specification asks for a function `sum_even(nums)` that returns the sum of even integers in the input list `nums`. The implementation is: ```python def sum_even(nums): return sum(n for n in nums if n % 2 == 0) ``` Let me check each test case: 1. **test_main**: `sum_even([...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00079
0.001
null
stop
1,218
386
Let me carefully analyze this code. The function `count_vowels(s)` is supposed to count the number of vowels in a string `s`, case-insensitive. Looking at the implementation: ```python def count_vowels(s): vowels = set('aeiou') text = s.lower() total = 0 index = 0 while index < len(text): ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00090
0.001
null
stop
1,094
399
Let me analyze this code carefully. Specification says: - `find_first_gt(nums, threshold)` should return the first element **strictly greater than** `threshold` - Or `None` if absent Implementation says: ```python def find_first_gt(nums, threshold): matches = [n for n in nums if n >= threshold] return matches...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00101
0.001
null
stop
1,555
427
Let me analyze this code carefully. The specification says: - Parse comma-separated integers - Handle optional surrounding whitespace - Empty segments must raise ValueError The implementation: ```python def parse_int_list(s): parts = tuple(part.strip() for part in s.split(', ')) if any(part == '' for part in ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00103
0.001
null
stop
893
405
Let me analyze this code carefully. The specification says: - Return the value with strictly highest count - If there's a tie for the top count, raise ValueError Let's trace through the implementation: ```python def majority_vote(labels): if not labels: raise ValueError('empty') unique = list(dict.fr...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00107
0.999
null
stop
1,096
359
Let me analyze this code carefully. The specification is to implement `sum_even(nums)` that returns the sum of even integers in `nums`. The implementation uses recursion with a helper function `subtotal(index)`. Let's trace through the logic: 1. `subtotal(index)` recursively processes the list starting at `index` 2...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00120
0.001
null
stop
949
399
Let me carefully analyze this implementation against the specification and tests. **Specification:** - Find the first element strictly greater than threshold - Return `None` if absent - "strictly greater than" means >, not >= **Implementation:** ```python def find_first_gt(nums, threshold): matches = [n for n in ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00121
0.001
null
stop
2,034
384
Let me analyze this code carefully. The specification asks for: - `unique_sorted(nums)` should return a sorted list of unique values from `nums` The implementation: ```python def unique_sorted(nums): remaining = list(nums) result = [] while remaining: smallest = min(remaining) result.appen...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00122
0.001
null
stop
2,865
384
Let me carefully analyze this implementation to see if it passes all tests. The specification asks for a function that: 1. Returns a sorted list 2. Contains unique values only 3. Values come from the input list `nums` Let's trace through the implementation: ```python def unique_sorted(nums): remaining = list(num...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00148
0.001
null
stop
1,105
389
Let me analyze this code carefully. The specification says: - Implement `rotate_left(lst, k)` returning a new list rotated left by `k` positions - Use `k % len(lst)` for non-empty lists The implementation: ```python def rotate_left(lst, k): if not lst: return [] i = k % len(lst) result = [] po...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00152
0.999
null
stop
1,240
360
Let me analyze this code carefully. The function `sum_even(nums)` should return the sum of all even integers in the input list. Let's trace through the implementation: ```python def sum_even(nums): def subtotal(index): if index == len(nums): return 0 n = nums[index] contributi...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00153
0.999
null
stop
1,265
406
Let me carefully analyze this code to determine if it will pass all tests. **Specification Analysis:** - Function should return the value with the strictly highest count - If there's a tie for the top count, should raise ValueError **Implementation Analysis:** ```python def majority_vote(labels): if not labels: ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00155
0.999
null
stop
1,378
406
Let me carefully analyze this code to determine if it passes all tests. **Specification:** - Return the value with strictly highest count - If there's a tie for the top count, raise ValueError **Implementation Analysis:** ```python def majority_vote(labels): if not labels: raise ValueError('empty') f...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00167
0.999
null
stop
1,170
416
Let me analyze this code carefully. The function `clamp(x, lo, hi)` should return `x` clamped into the interval `[lo, hi]`, meaning: - If x < lo, return lo - If x > hi, return hi - Otherwise (lo <= x <= hi), return x The implementation is: ```python def clamp(x, lo, hi): cases = ((x < lo, lo), (x > hi, hi)) f...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00170
0.001
null
stop
1,125
405
Let me analyze this code carefully. The specification says: - Return the value with strictly highest count - If the top count is tied, raise ValueError Let's trace through the implementation: ```python def majority_vote(labels): if not labels: raise ValueError('empty') unique = list(dict.fromkeys(lab...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00172
0.001
null
stop
1,211
389
Let me analyze this code carefully. The specification says: - Find the first element strictly greater than `threshold` - Return that element or `None` if absent The implementation: ```python def find_first_gt(nums, threshold): for _, n in enumerate(nums): if n < threshold: # This looks suspicious ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00174
0.999
null
stop
860
364
Let me carefully analyze this code to determine if it will pass all tests. **Specification Analysis:** - Function name: `second_largest` - Input: `nums` (list of numbers) - Output: Second largest distinct value in `nums` - Exception: Raise ValueError if all elements are equal (no distinct second largest exists) **Imp...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00178
0.999
null
stop
1,142
387
Let me analyze this code carefully. The function `unique_sorted` should: 1. Remove duplicates from the input list 2. Return a sorted list of unique values Let's trace through the implementation: ```python def unique_sorted(nums): remaining = set(nums) # Convert to set, which automatically removes duplicates ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00189
0.999
null
stop
1,087
382
Let me carefully analyze this code. The specification says: - `clamp(x, lo, hi)` should return `x` clamped into the closed interval `[lo, hi]` - Assume `lo <= hi` The implementation: ```python def clamp(x, lo, hi): cases = ((x < lo, lo), (x > hi, hi)) for applies, value in cases: if applies: ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00209
0.001
null
stop
1,353
384
Let me carefully analyze this code to see if it will pass all the tests. The function is supposed to rotate a list left by k positions. Let's trace through the logic: 1. `if not lst:` - handles empty list, returns `[]`. This is correct. 2. `i = k % len(lst)` - calculates the effective rotation (handles k being larg...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00222
0.001
null
stop
2,113
435
Let me carefully analyze this code to determine if it passes all tests. ## Specification Analysis - Parse comma-separated integers - Optional surrounding whitespace around each segment - Empty segments must raise ValueError ## Implementation Analysis ```python def parse_int_list(s): parts = tuple(part.strip() fo...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00260
0.001
null
stop
940
399
Let me analyze this code carefully. The specification says: - `find_first_gt(nums, threshold)` should return the first element **strictly greater than** `threshold` - Return `None` if no such element exists The implementation is: ```python def find_first_gt(nums, threshold): matches = [n for n in nums if n >= thr...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00280
0.001
null
stop
1,457
390
Let me carefully analyze this code to see if it passes all tests. The function `rotate_left(lst, k)` should rotate a list left by `k` positions. Let's trace through the logic: 1. First, it checks if the list is empty. If so, returns an empty list. This handles the test case `test_empty`. 2. For non-empty lists: ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00292
0.999
null
stop
772
318
Let me analyze this code carefully. The specification asks for a function `median3(a, b, c)` that returns the median of three distinct integers. The implementation is: ```python def median3(a, b, c): return sorted([a, b, c])[1] ``` Let's trace through what this does: 1. It creates a list with the three values: [...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00309
0.001
null
stop
1,173
381
Let me carefully analyze this code. **Specification:** - Parse comma-separated integers with optional surrounding whitespace - Empty segments must raise ValueError **Implementation Analysis:** ```python def parse_int_list(s): parts = list(map(str.strip, s.split(','))) result = [] for part in parts: ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00313
0.001
null
stop
1,150
364
Let me analyze this code carefully. **Specification:** - Implement `second_largest(nums)` - Return the second largest DISTINCT value - If all elements are equal (no distinct second largest exists), raise ValueError **Implementation:** ```python def second_largest(nums): values = sorted(nums) if len(values) < ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
ex00327
0.999
null
stop
1,199
416
Let me analyze this implementation carefully. The spec says: `clamp(x, lo, hi)` should return `x` clamped into the closed interval `[lo, hi]`. The implementation: ```python def clamp(x, lo, hi): cases = ((x < lo, lo), (x > hi, hi)) for applies, value in cases: if applies: return value ...
[ { "role": "user", "content": "You are an expert code reviewer. You will be shown a Python specification, implementation, and test suite. Your job is to predict the probability that the implementation passes ALL tests when executed.\n\nCarefully analyze the code for common bugs: off-by-one errors, missing ed...
End of preview. Expand in Data Studio

speccast-code-correctness-v1-2-blind-val-predictions

Blind validation (n=975) predictions from final SpecCast v1.2 LoRA. Labels withheld.

  • Base model: Qwen/Qwen3.5-4B
  • Adapter: LightningRodLabs/speccast-code-correctness-v1-2-qwen3.5-4b-lora
  • Source benchmark: aligncast/speccast-benchmark @ 7cef3c2…
  • Rows: 975
  • Upstream artifact: LightningRodLabs/speccast-code-correctness-v1-2-blind-val-qwen3-5-4b-blind-val-ckpt-20260722T122146Z

Blind splits keep labels empty by design.

Downloads last month
32