--- dataset_info: - config_name: Chinese features: - name: file_id dtype: string - name: content dtype: string - name: repo dtype: string - name: path dtype: string - name: original_comment dtype: string - name: masked_data_Qwen/CodeQwen1.5-7B dtype: string - name: predict_Qwen/CodeQwen1.5-7B dtype: string - name: predicted_comment_Qwen/CodeQwen1.5-7B dtype: string - name: masked_data_bigcode/starcoder2-7b dtype: string - name: predict_bigcode/starcoder2-7b dtype: string - name: predicted_comment_bigcode/starcoder2-7b dtype: string - name: masked_data_ibm-granite/granite-8b-code-base dtype: string - name: predict_ibm-granite/granite-8b-code-base dtype: string - name: predicted_comment_ibm-granite/granite-8b-code-base dtype: string - name: masked_data_meta-llama/CodeLlama-7b-hf dtype: string - name: predict_meta-llama/CodeLlama-7b-hf dtype: string - name: predicted_comment_meta-llama/CodeLlama-7b-hf dtype: string - name: masked_data_google/codegemma-7b dtype: string - name: predict_google/codegemma-7b dtype: string - name: predicted_comment_google/codegemma-7b dtype: string - name: expert_accuracy_Qwen/CodeQwen1.5-7B dtype: string - name: error_codes_Qwen/CodeQwen1.5-7B dtype: string - name: expert_accuracy_bigcode/starcoder2-7b dtype: string - name: error_codes_bigcode/starcoder2-7b dtype: string - name: expert_accuracy_google/codegemma-7b dtype: string - name: error_codes_google/codegemma-7b dtype: string - name: expert_accuracy_ibm-granite/granite-8b-code-base dtype: string - name: error_codes_ibm-granite/granite-8b-code-base dtype: string - name: expert_accuracy_meta-llama/CodeLlama-7b-hf dtype: string - name: error_codes_meta-llama/CodeLlama-7b-hf dtype: string splits: - name: train num_bytes: 21709444 num_examples: 500 download_size: 9007265 dataset_size: 21709444 - config_name: Dutch features: - name: file_id dtype: string - name: content dtype: string - name: repo dtype: string - name: path dtype: string - name: original_comment dtype: string - name: masked_data_Qwen/CodeQwen1.5-7B dtype: string - name: predict_Qwen/CodeQwen1.5-7B dtype: string - name: predicted_comment_Qwen/CodeQwen1.5-7B dtype: string - name: masked_data_bigcode/starcoder2-7b dtype: string - name: expert_accuracy_Qwen/CodeQwen1.5-7B dtype: string - name: error_codes_Qwen/CodeQwen1.5-7B dtype: string - name: predict_bigcode/starcoder2-7b dtype: string - name: predicted_comment_bigcode/starcoder2-7b dtype: string - name: masked_data_ibm-granite/granite-8b-code-base dtype: string - name: expert_accuracy_bigcode/starcoder2-7b dtype: string - name: error_codes_bigcode/starcoder2-7b dtype: string - name: predict_ibm-granite/granite-8b-code-base dtype: string - name: predicted_comment_ibm-granite/granite-8b-code-base dtype: string - name: masked_data_meta-llama/CodeLlama-7b-hf dtype: string - name: expert_accuracy_ibm-granite/granite-8b-code-base dtype: string - name: error_codes_ibm-granite/granite-8b-code-base dtype: string - name: predict_meta-llama/CodeLlama-7b-hf dtype: string - name: predicted_comment_meta-llama/CodeLlama-7b-hf dtype: string - name: masked_data_google/codegemma-7b dtype: string - name: expert_accuracy_meta-llama/CodeLlama-7b-hf dtype: string - name: error_codes_meta-llama/CodeLlama-7b-hf dtype: string - name: predict_google/codegemma-7b dtype: string - name: predicted_comment_google/codegemma-7b dtype: string - name: expert_accuracy_google/codegemma-7b dtype: string - name: error_codes_google/codegemma-7b dtype: string splits: - name: train num_bytes: 24547575 num_examples: 500 download_size: 9149540 dataset_size: 24547575 - config_name: English features: - name: file_id dtype: string - name: content dtype: string - name: repo dtype: string - name: path dtype: string - name: original_comment dtype: string - name: masked_data_Qwen/CodeQwen1.5-7B dtype: string - name: predict_Qwen/CodeQwen1.5-7B dtype: string - name: predicted_comment_Qwen/CodeQwen1.5-7B dtype: string - name: masked_data_bigcode/starcoder2-7b dtype: string - name: predict_bigcode/starcoder2-7b dtype: string - name: predicted_comment_bigcode/starcoder2-7b dtype: string - name: masked_data_ibm-granite/granite-8b-code-base dtype: string - name: predict_ibm-granite/granite-8b-code-base dtype: string - name: predicted_comment_ibm-granite/granite-8b-code-base dtype: string - name: masked_data_meta-llama/CodeLlama-7b-hf dtype: string - name: predict_meta-llama/CodeLlama-7b-hf dtype: string - name: predicted_comment_meta-llama/CodeLlama-7b-hf dtype: string - name: masked_data_google/codegemma-7b dtype: string - name: predict_google/codegemma-7b dtype: string - name: predicted_comment_google/codegemma-7b dtype: string - name: error_codes_Qwen/CodeQwen1.5-7B dtype: string - name: expert_accuracy_Qwen/CodeQwen1.5-7B dtype: string - name: error_codes_bigcode/starcoder2-7b dtype: string - name: expert_accuracy_bigcode/starcoder2-7b dtype: string - name: error_codes_ibm-granite/granite-8b-code-base dtype: string - name: expert_accuracy_ibm-granite/granite-8b-code-base dtype: string - name: error_codes_meta-llama/CodeLlama-7b-hf dtype: string - name: expert_accuracy_meta-llama/CodeLlama-7b-hf dtype: string - name: error_codes_google/codegemma-7b dtype: string - name: expert_accuracy_google/codegemma-7b dtype: string splits: - name: train num_bytes: 20576413 num_examples: 500 download_size: 8169212 dataset_size: 20576413 - config_name: Greek features: - name: file_id dtype: string - name: content dtype: string - name: repo dtype: string - name: path dtype: string - name: original_comment dtype: string - name: masked_data_Qwen/CodeQwen1.5-7B dtype: string - name: predict_Qwen/CodeQwen1.5-7B dtype: string - name: predicted_comment_Qwen/CodeQwen1.5-7B dtype: string - name: masked_data_bigcode/starcoder2-7b dtype: string - name: predict_bigcode/starcoder2-7b dtype: string - name: predicted_comment_bigcode/starcoder2-7b dtype: string - name: masked_data_ibm-granite/granite-8b-code-base dtype: string - name: predict_ibm-granite/granite-8b-code-base dtype: string - name: predicted_comment_ibm-granite/granite-8b-code-base dtype: string - name: masked_data_meta-llama/CodeLlama-7b-hf dtype: string - name: predict_meta-llama/CodeLlama-7b-hf dtype: string - name: predicted_comment_meta-llama/CodeLlama-7b-hf dtype: string - name: masked_data_google/codegemma-7b dtype: string - name: predict_google/codegemma-7b dtype: string - name: predicted_comment_google/codegemma-7b dtype: string - name: error_codes_bigcode/starcoder2-7b dtype: string - name: error_codes_ibm-granite/granite-8b-code-base dtype: string - name: error_codes_meta-llama/CodeLlama-7b-hf dtype: string - name: error_codes_google/codegemma-7b dtype: string - name: error_codes_Qwen/CodeQwen1.5-7B dtype: string - name: expert_accuracy_bigcode/starcoder2-7b dtype: string - name: expert_accuracy_ibm-granite/granite-8b-code-base dtype: string - name: expert_accuracy_meta-llama/CodeLlama-7b-hf dtype: string - name: expert_accuracy_google/codegemma-7b dtype: string - name: expert_accuracy_Qwen/CodeQwen1.5-7B dtype: string splits: - name: train num_bytes: 25228509 num_examples: 500 download_size: 8978147 dataset_size: 25228509 - config_name: Polish features: - name: file_id dtype: string - name: repo dtype: string - name: path dtype: string - name: content dtype: string - name: original_comment dtype: string - name: masked_data_Qwen/CodeQwen1.5-7B dtype: string - name: predict_Qwen/CodeQwen1.5-7B dtype: string - name: predicted_comment_Qwen/CodeQwen1.5-7B dtype: string - name: masked_data_bigcode/starcoder2-7b dtype: string - name: predict_bigcode/starcoder2-7b dtype: string - name: predicted_comment_bigcode/starcoder2-7b dtype: string - name: masked_data_ibm-granite/granite-8b-code-base dtype: string - name: predict_ibm-granite/granite-8b-code-base dtype: string - name: predicted_comment_ibm-granite/granite-8b-code-base dtype: string - name: masked_data_meta-llama/CodeLlama-7b-hf dtype: string - name: predict_meta-llama/CodeLlama-7b-hf dtype: string - name: predicted_comment_meta-llama/CodeLlama-7b-hf dtype: string - name: masked_data_google/codegemma-7b dtype: string - name: predict_google/codegemma-7b dtype: string - name: predicted_comment_google/codegemma-7b dtype: string - name: error_codes_Qwen/CodeQwen1.5-7B dtype: string - name: expert_accuracy_Qwen/CodeQwen1.5-7B dtype: string - name: error_codes_bigcode/starcoder2-7b dtype: string - name: expert_accuracy_bigcode/starcoder2-7b dtype: string - name: error_codes_ibm-granite/granite-8b-code-base dtype: string - name: expert_accuracy_ibm-granite/granite-8b-code-base dtype: string - name: error_codes_meta-llama/CodeLlama-7b-hf dtype: string - name: expert_accuracy_meta-llama/CodeLlama-7b-hf dtype: string - name: error_codes_google/codegemma-7b dtype: string - name: expert_accuracy_google/codegemma-7b dtype: string splits: - name: train num_bytes: 17812544 num_examples: 500 download_size: 7250273 dataset_size: 17812544 configs: - config_name: Chinese data_files: - split: train path: Chinese/train-* - config_name: Dutch data_files: - split: train path: Dutch/train-* - config_name: English data_files: - split: train path: English/train-* - config_name: Greek data_files: - split: train path: Greek/train-* - config_name: Polish data_files: - split: train path: Polish/train-* --- # Fixed-8 Based on fixed-7 revision `14e85fe00a8b284cd226c58281ddd8e6b990b190`. Replaces six Greek rows with missing expert labels with six newly labelled samples. All five language configurations retain 500 training rows (2,500 total). All other rows are unchanged. | Removed ID | Replacement ID | |---|---| | 8000_5 | 1056_0 | | 8000_14 | 4357_8 | | 8000_15 | 4848_9 | | 8000_16 | 29069_13 | | 8000_17 | 1385_4 | | 8000_18 | 5142_0 | All 500 Greek rows now have all five expert accuracy labels. Original supplied predictions and error-code strings are preserved; blank error-code fields are retained as provided. This release does not remove pre-existing file reuse or certify absence of partial code clones. Scores and judge results from fixed-7 must be recomputed for the six replacement IDs. ## CodeGemma comment extraction correction Corrected from dataset revision `292880b74aa2ec58dc1b00524a70b060359bf0fa`. Only `predicted_comment_google/codegemma-7b` was changed: each affected completion is truncated immediately before its first `<|file_separator|>` token. The token and all subsequent generated text are removed; the preceding text and whitespace are preserved exactly. Raw `predict_google/codegemma-7b` generations, prompts, reference comments, expert labels, error codes, row order, and every other generator's columns are unchanged. | Language | Corrected comments | |---|---:| | Chinese | 464 | | Dutch | 480 | | English | 477 | | Greek | 469 | | Polish | 469 | 11 affected completions contain only whitespace or no text before the boundary; these are preserved as such, with their supplied expert labels unchanged. Previously computed neural scores for changed comments must be recomputed from these corrected inputs. ## Remaining comment extraction corrections Corrected from revision `d02cf177485908c0d26e71c26f4a584c0e38dd5d`, preserving the preceding CodeGemma correction. Only the extracted comment columns for CodeLlama, CodeQwen 1.5, and GraniteCode change. Fourteen comments are truncated before their first generation-control token: six CodeLlama `` endings, two CodeLlama `` continuations, five GraniteCode `<|endoftext|>` endings, and one CodeQwen `<|endoftext|>` ending. The preceding text and whitespace are preserved exactly. For Greek file `777248_1`, CodeQwen's extracted-comment field incorrectly equals its entire raw generation. Its exact stored `masked_data` prompt prefix (791 characters, ending in ``) is removed, preserving the 1,378-character generated completion exactly. | Language | CodeLlama | CodeQwen 1.5 | GraniteCode | |---|---:|---:|---:| | English | 2 | 0 | 0 | | Greek | 6 | 2 | 5 | These 15 corrections introduce no new blank completions. Raw generations, prompts, references, expert labels, error codes, row order, CodeGemma, and StarCoder2 values are unchanged. Previously computed neural scores for changed comments must be recomputed.