Update dataset paths and output path

harness_eval.slurm

@@ -57,7 +57,7 @@ lm_eval \
 --tasks $tasks \
 --num_fewshot 0 \
 --batch_size 8 \
---output_path evals \
+--output_path eval_results \
 --write_out \
 --log_samples \
 --verbosity DEBUG \

lm_eval/tasks/pega/default/_default_template_yaml

-dataset_path: yentinglin/legal_benchmark
+dataset_path: yentinglin/PegaEval
 test_split: train
 output_type: multiple_choice
 process_docs: !function utils.process_docs
@@ -9,8 +9,5 @@ metric_list:
   - metric: acc
     aggregation: mean
     higher_is_better: true
-  - metric: acc_norm
-    aggregation: mean
-    higher_is_better: true
 metadata:
   version: 0.1

lm_eval/tasks/tmlu/default/_default_template_yaml

@@ -12,8 +12,5 @@ metric_list:
   - metric: acc
     aggregation: mean
     higher_is_better: true
-  - metric: acc_norm
-    aggregation: mean
-    higher_is_better: true
 metadata:
   version: 0.1

lm_eval/tasks/tmmluplus/default/_default_template_yaml

@@ -12,8 +12,5 @@ metric_list:
   - metric: acc
     aggregation: mean
     higher_is_better: true
-  - metric: acc_norm
-    aggregation: mean
-    higher_is_better: true
 metadata:
   version: 0.1

lm_eval/tasks/tw_legal/default/_default_template_yaml

-dataset_path: yentinglin/PegaEval
+dataset_path: lianghsun/tw-legal-benchmark-v1
 test_split: train
 output_type: multiple_choice
 process_docs: !function utils.process_docs
@@ -9,8 +9,5 @@ metric_list:
   - metric: acc
     aggregation: mean
     higher_is_better: true
-  - metric: acc_norm
-    aggregation: mean
-    higher_is_better: true
 metadata:
   version: 0.1

lm_eval/tasks/twllm_eval/default/_default_template_yaml

@@ -9,8 +9,5 @@ metric_list:
   - metric: acc
     aggregation: mean
     higher_is_better: true
-  - metric: acc_norm
-    aggregation: mean
-    higher_is_better: true
 metadata:
   version: 0.1