From 2116c73ab3212359ce5f099e6de3aa59ce0ed69f Mon Sep 17 00:00:00 2001 From: komaksym Date: Sat, 16 Aug 2025 18:15:13 +0200 Subject: [PATCH 1/2] fix chunk calculation --- .../learn.md | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/questions/110_evaluate-translation-quality-with-meteor-score/learn.md b/questions/110_evaluate-translation-quality-with-meteor-score/learn.md index 1a981bbd..992c97b7 100644 --- a/questions/110_evaluate-translation-quality-with-meteor-score/learn.md +++ b/questions/110_evaluate-translation-quality-with-meteor-score/learn.md @@ -66,10 +66,9 @@ F_mean = (Precision * Recall) / ### 5. Chunk Calculation - Contiguous matched sequences: - 1. ['quick', 'brown', 'fox'] - 2. ['jumps', 'over'] - 3. ['lazy', 'dog'] -- Number of Chunks: 3 + 1. ['quick', 'brown', 'fox', jumps', 'over'] + 2. ['lazy', 'dog'] +- Number of Chunks: 2 - Total Number of Unigram Matches: 7 ### 6. Penalty Calculation (betta = 3, gamma = 0.5) From 2e52860eaa8a5320d070798584a7ff676345df6a Mon Sep 17 00:00:00 2001 From: komaksym Date: Sat, 16 Aug 2025 18:16:09 +0200 Subject: [PATCH 2/2] fix unigram reasoning --- .../example.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/questions/110_evaluate-translation-quality-with-meteor-score/example.json b/questions/110_evaluate-translation-quality-with-meteor-score/example.json index d1084515..26203c35 100644 --- a/questions/110_evaluate-translation-quality-with-meteor-score/example.json +++ b/questions/110_evaluate-translation-quality-with-meteor-score/example.json @@ -1,5 +1,5 @@ { "input": "meteor_score('Rain falls gently from the sky', 'Gentle rain drops from the sky')", "output": "0.625", - "reasoning": "The function identifies 4 unigram matches ('rain', 'gently'/'gentle', 'from', 'sky'), computes precision (4/6) and recall (4/5), calculates an F-mean, and then apply a small penalty for two chunks." + "reasoning": "The function identifies 4 unigram matches ('rain', 'from', 'the', 'sky'), computes precision (4/6) and recall (4/6), calculates an F-mean, and then apply a small penalty for two chunks." }