fix do_sample

lm_eval/tasks/longbench/passage_retrieval_en_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 32
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.retrieval_score
     aggregation: mean

lm_eval/tasks/longbench/passage_retrieval_zh.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 32
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.retrieval_zh_score
     aggregation: mean

lm_eval/tasks/longbench/qasper.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 128
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.qa_f1_score
     aggregation: mean

lm_eval/tasks/longbench/qasper_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 128
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.qa_f1_score
     aggregation: mean

lm_eval/tasks/longbench/qmsum.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 512
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.rouge_score
     aggregation: mean

lm_eval/tasks/longbench/repobench-p.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 64
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.code_sim_score
     aggregation: mean

lm_eval/tasks/longbench/repobench-p_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 64
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.code_sim_score
     aggregation: mean

lm_eval/tasks/longbench/samsum.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 128
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.rouge_score
     aggregation: mean

lm_eval/tasks/longbench/samsum_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 128
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.rouge_score
     aggregation: mean

lm_eval/tasks/longbench/trec.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 64
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.classification_score
     aggregation: mean

lm_eval/tasks/longbench/trec_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 64
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.classification_score
     aggregation: mean

lm_eval/tasks/longbench/triviaqa.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 32
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.qa_f1_score
     aggregation: mean

lm_eval/tasks/longbench/triviaqa_e.yaml

@@ -10,7 +10,7 @@ doc_to_target: '{{answers}}'
 generation_kwargs:
   max_gen_toks: 32
   temperature: 1
-  do_sample: False
+  do_sample: True
 metric_list:
   - metric: !function metrics.qa_f1_score
     aggregation: mean