diff --git a/flutter/cpp/datasets/ifeval.cc b/flutter/cpp/datasets/ifeval.cc
index 83f950623..07834ebbd 100644
--- a/flutter/cpp/datasets/ifeval.cc
+++ b/flutter/cpp/datasets/ifeval.cc
@@ -77,8 +77,6 @@ std::vector<uint8_t> IFEval::ProcessOutput(const int sample_idx,
   sample_output_tokens_[sample_idx] = output_tokens;
   used_sample_ids_.insert(sample_idx);
 
-  LOG(INFO) << "Processed " << std::to_string(used_sample_ids_.size()) << "/29";
-
   return {1};
 }
 
diff --git a/flutter/cpp/datasets/ifeval.h b/flutter/cpp/datasets/ifeval.h
index 7a3f15d21..5ce4ea357 100644
--- a/flutter/cpp/datasets/ifeval.h
+++ b/flutter/cpp/datasets/ifeval.h
@@ -46,7 +46,7 @@ class IFEval : public Dataset {
 
   size_t TotalSampleCount() override { return samples_.size(); }
 
-  size_t PerformanceSampleCount() override { return 1; }
+  size_t PerformanceSampleCount() override { return samples_.size(); }
 
   void LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) override;
 
diff --git a/flutter/cpp/datasets/ifeval_utils/common.h b/flutter/cpp/datasets/ifeval_utils/common.h
index 0345d0fbf..629a55575 100644
--- a/flutter/cpp/datasets/ifeval_utils/common.h
+++ b/flutter/cpp/datasets/ifeval_utils/common.h
@@ -51,8 +51,6 @@ inline bool contains_string(const std::string& text,
 inline bool contains_word(const std::string& text, const std::string& word) {
   if (word.empty()) return false;
 
-  LOG(INFO) << "searching for '" << word << "'...";
-
   auto to_lower_ascii = [](std::string s) {
     for (char& c : s) c = std::tolower(static_cast<unsigned char>(c));
     return s;
diff --git a/flutter/cpp/datasets/mmlu_gen.cc b/flutter/cpp/datasets/mmlu_gen.cc
index 921d47714..4b79efc44 100644
--- a/flutter/cpp/datasets/mmlu_gen.cc
+++ b/flutter/cpp/datasets/mmlu_gen.cc
@@ -7,8 +7,6 @@
 #include "tensorflow/core/example/example.pb.h"
 #include "tensorflow/core/example/feature_util.h"
 
-#define li LOG(INFO) << "li:" << __FILE__ << ":" << __LINE__ << "@" << __func__
-
 namespace mlperf {
 namespace mobile {
 
@@ -18,8 +16,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
   sp_processor = std::unique_ptr<sentencepiece::SentencePieceProcessor>(
       LoadSentencePieceProcessor(sp_path));
 
-  li;
-
   // Load all TFRecord samples into memory
   // NOTE this can be moved to LoadSamplesToRam, but will cause delays between
   // queries due to IO reads happening between them
@@ -41,8 +37,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
       std::string preface = input.substr(0, input.find("\n\n") + 2);
 
       input = preface + "Question: " + question_formatted;
-
-      LOG(INFO) << input;
     }
 
     // std::string input_formatted = FormatLlamaUserPrompt(input, "Provide only
@@ -59,7 +53,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
     samples_.push_back(std::move(sample));
     sample_output_tokens_.push_back(std::vector<int>());
   }
-  li;
 }
 
 void MmluGen::LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) {
@@ -78,8 +71,6 @@ void MmluGen::UnloadSamplesFromRam(
 std::vector<void*> MmluGen::GetData(int sample_idx) {
   std::vector<void*> data;
 
-  li;
-  LOG(INFO) << "Sample ID: " << std::to_string(sample_idx);
   if (sample_idx < samples_.size()) {
     data.push_back(reinterpret_cast<void*>(
         const_cast<std::vector<int>*>(&(samples_[sample_idx]->input_tokens))));
@@ -92,18 +83,12 @@ std::vector<uint8_t> MmluGen::ProcessOutput(const int sample_idx,
                                             const std::vector<void*>& outputs) {
   if (sample_idx >= samples_.size() || outputs.empty()) return {0};
 
-  li;
   const auto& output_tokens =
       *(reinterpret_cast<std::vector<int>*>(outputs[0]));
 
-  li;
   sample_output_tokens_[sample_idx] = output_tokens;
   used_sample_ids_.insert(sample_idx);
 
-  li;
-  LOG(INFO) << "Processed " << std::to_string(used_sample_ids_.size())
-            << "/100";
-
   return {1};
 }
 
@@ -117,14 +102,9 @@ bool MmluGen::ComputeSampleAccuracy(const int sample_idx) {
   std::string prediction;
   sp_processor->Decode(sample_output_tokens_[sample_idx], &prediction).ok();
 
-  li;
-  LOG(INFO) << "index: " << std::to_string(sample_idx) << std::endl;
-  LOG(INFO) << "Output: [[[" << prediction << "]]]" << std::endl;
-
   char predicted_char = find_answer_char(prediction);
   const std::string& correct = samples_[sample_idx]->answer;
 
-  li;
   return (predicted_char == correct[0]);
 }
 
diff --git a/flutter/cpp/datasets/mmlu_gen.h b/flutter/cpp/datasets/mmlu_gen.h
index 72f6c6a71..18ecc4d0d 100644
--- a/flutter/cpp/datasets/mmlu_gen.h
+++ b/flutter/cpp/datasets/mmlu_gen.h
@@ -26,7 +26,7 @@ class MmluGen : public Dataset {
 
   size_t TotalSampleCount() override { return samples_.size(); }
 
-  size_t PerformanceSampleCount() override { return 1; }
+  size_t PerformanceSampleCount() override { return samples_.size(); }
 
   void LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) override;