update

lm_eval/tasks/squadv2/README.md

@@ -2,25 +2,44 @@
 
 ### Paper
 
-Title: `paper title goes here`
-Abstract: `link to paper PDF or arXiv abstract goes here`
+Title: `Know What You Don’t Know: Unanswerable Questions for SQuAD`
+Abstract: https://arxiv.org/abs/1806.03822
 
-`Short description of paper / benchmark goes here:`
+Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset,
+consisting of questions posed by crowdworkers on a set of Wikipedia articles,
+where the answer to every question is a segment of text, or span, from the
+corresponding reading passage, or the question might be unanswerable.
+SQuAD2.0 combines the 100,000 questions in SQuAD1.1 with over 50,000 unanswerable
+questions written adversarially by crowdworkers to look similar to answerable ones.
+To do well on SQuAD2.0, systems must not only answer questions when possible, but
+also determine when no answer is supported by the paragraph and abstain from answering.
 
-Homepage: `homepage to the benchmark's website goes here, if applicable`
+Homepage: https://rajpurkar.github.io/SQuAD-explorer/
 
 
 ### Citation
 
 ```
-BibTeX-formatted citation goes here
+@misc{rajpurkar2018know,
+    title={Know What You Don't Know: Unanswerable Questions for SQuAD},
+    author={Pranav Rajpurkar and Robin Jia and Percy Liang},
+    year={2018},
+    eprint={1806.03822},
+    archivePrefix={arXiv},
+    primaryClass={cs.CL}
+}
 ```
 
-### Subtasks
+### Groups and Tasks
 
-List or describe tasks defined in this folder, and their names here:
-* `task_name`: `1-sentence description of what this particular task does`
-* `task_name2`: .....
+#### Groups
+
+* `squadv2_complete`: Runs both `squadv2` and `squadv2_noans_loglikelihood`
+
+#### Tasks
+
+* `squadv2`: `Default squadv2 task`
+* `squadv2_noans_loglikelihood`: `Additional task to acquire the probability of model predicting there is no answer`
 
 ### Checklist
 

lm_eval/tasks/squadv2/default.yaml

+include: _template_yaml
 task: squadv2
-dataset_path: squad_v2
 output_type: greedy_until
-training_split: train
-validation_split: validation
-doc_to_text: "Title: {{title}}\n\nBackground: {{context}}\n\nQuestion: {{question}}\n\n Answer:"
-doc_to_target: "{% if answers.text| length > 0 %}{{answers.text}}{% else %}{{['']}}{% endif %}"
-target_delimiter: ""
-should_decontaminate: true
-doc_to_decontamination_query: context
 generation_kwargs:
   until:
     - "\n"
-# filter_list:
-#   - name: remove_whitespace
-#     filter:
-#       - function: remove_whitespace
-#       - function: take_first
 metric_list:
   - metric: !function utils.exact
     aggregation: mean

lm_eval/tasks/squadv2/no_ans.yaml

-include: default.yaml
+include: _template_yaml
 task: squadv2_noans_loglikelihood
-dataset_path: squad_v2
 output_type: loglikelihood
-training_split: train
-validation_split: validation
 doc_to_target: " unanswerable"
 metric_list:
   - metric: perplexity