AfroBench: How Good are Large Language Models on African Languages? (#2825)

* add afrixnli to task

* add chat completion

* remove chat completion -untested

* afrimmlu added

* afrimmlu folder update

* afrimmlu folder update

* updated prompt

* remove print

* add afrimgsm -direct

* add squad metric

* fix bash script

* remove direct util, update common yaml

* remove print

* add few show. metric fixes

* fix direct path, add bash script for gpt models

* added transate test

* update afrixnli tasks

* update afrixnli tasks

* update metrics for afrixnli

* prompt translations fix

* prompt translations fix

* filter and metric fix -mgsm

* remove squad metric

* remove squad metric

* add f1 score to mgsm

* add f1 score to mgsm

* update native-direct with lin

* change f1 function

* add lin to utils

* add utils

* remove test limit

* remove test configs

* add swahili to mmlu

* change eng to ewe in ewe yaml mmlu

* add squad metric to mgsm, remove whitespace filter

* added translate test

* added afrixnli_translate

* fix exact match valueError

* fix exact match valueError

* restructure mmlu folder

* spacing

* remove afrimmlu_translate folder

* add utility

* format task name, clean ups

* modefied mgsm

* update on afrimgsm

* update on afrimgsm

* removed utils

* other mgsm varieties

* other mgsm varieties

* adding trasnslate direct

* Update translate_direct_yaml

* add manual xnli prompt, add multichoice for openai models, and adapt multichoice metric for openai model

* edit for open models

* Update translate_direct_yaml

* add verbalizer for xnli

* change xnli from multiple choice to generate

* add manual accuracy scores

* revert xnli to multiple choice

* change afrimgsm utils

* revert xnli to multiple_choice

* cleanups and readmes

* remove openai fixes and unused regex

* pr review changes

* revert metrics.py, task.py and extraction.py to main version

* add afrisenti

* utilities

* pulled from main

* add afrixnli

* add afrimmlu

* update afrixnli prompts

* mising senti language

* fix afrisenti prompt 2

* fix afrisenti prompts

* fix afrisenti prompts

* configure task grouping

* add multiple prompts to afrixnli for irokobench

* add multiple prompts to afrimmlu for irokobench

* Update afrixnli_yaml

* fixes and moves

* fixes and moves

* afrimmlu multiple prompts configs

* remove validation set from afrimmlu

* remove eng from afrimmlu translate test

* correct dataset path

* multiple prompts for mgsm

* file restructure

* afribench grouping

* repo restructuring

* repo restructuring

* update exact match to hugging face exact match and add new mgsm language

* remove decontamination

* update generation kwargs

* update generation kwargs for all mgsm prompts

* remove lang

* update generation kwargs for afrimgsm translatetest

* add afrimgsm cot for direct and translate

* remove eng from translate-cot

* add masakhaPOS tasks

* remove changes from task script

* add masakhanews tasks

* add uhura arc easy

* add afriqa and belebele files

* add tags for easier run. add naija rc

* add new metrics and transformation scripts

* fix afriqa swa fewshot split

* add naijarc

* add afrobench lite tasks

* update afrobench

* update afrobench

* remove unverified files to avoid bugs

* remove files not needed

* add afrobench tasks

* add afrobench tasks

* change to version 1

* change to version 1

* update afrobench

* update afrobench

* restore metric to original script

* update readme instructions

* add individual dataset readmes

* add link to collections

* correct run script

* align with main

* align with main

* align with main

* align with main

* align with main

* align with main

* align with main

* align with main

* failed run fixes

* failed run fixes

* add afrimgsm cot

* Apply precommit fixes

* update mafand dataset name

* pull request fixes

* remove afrihate due to availability

---------
Co-authored-by: Israel Abebe Azime <azime@cg.uni-saarland.de>
Co-authored-by: Israel Abebe Azime <se.israel.abebe@gmail.com>
Co-authored-by: David Adelani <davlanade@gmail.com>
Co-authored-by: theyorubayesian <akin.o.oladipo@gmail.com>

lm_eval/tasks/afrixnli/translate/prompt_3/afrixnli_translate_wol.yaml

+# Generated by utils.py
+dataset_name: wol
+doc_to_text: "Given the following premise and hypothesis in Wolof, identify if the\
+  \ premise entails, contradicts, or is neutral towards the hypothesis. Please respond\
+  \ with exact 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_wol_prompt_3

lm_eval/tasks/afrixnli/translate/prompt_3/afrixnli_translate_xho.yaml

+# Generated by utils.py
+dataset_name: xho
+doc_to_text: "Given the following premise and hypothesis in isiXhosa, identify if\
+  \ the premise entails, contradicts, or is neutral towards the hypothesis. Please\
+  \ respond with exact 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_xho_prompt_3

lm_eval/tasks/afrixnli/translate/prompt_3/afrixnli_translate_yaml

+tag: afrixnli_tt_tasks
+dataset_path: masakhane/afrixnli-translate-test
+dataset_name: null
+output_type: multiple_choice
+test_split: test
+fewshot_split: test
+doc_to_target: !function utils.doc_to_target
+doc_to_choice:
+  - "entailment"
+  - "neutral"
+  - "contradiction"
+should_decontaminate: true
+doc_to_decontamination_query: premise
+metric_list:
+  - metric: f1
+    aggregation: !function utils.weighted_f1_score
+    average: weighted
+    higher_is_better: True
+    ignore_case: true
+    ignore_punctuation: true
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+    ignore_case: true
+    ignore_punctuation: true
+metadata:
+  version: 1.0

lm_eval/tasks/afrixnli/translate/prompt_3/afrixnli_translate_yor.yaml

+# Generated by utils.py
+dataset_name: yor
+doc_to_text: "Given the following premise and hypothesis in Yoruba, identify if the\
+  \ premise entails, contradicts, or is neutral towards the hypothesis. Please respond\
+  \ with exact 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_yor_prompt_3

lm_eval/tasks/afrixnli/translate/prompt_3/afrixnli_translate_zul.yaml

+# Generated by utils.py
+dataset_name: zul
+doc_to_text: "Given the following premise and hypothesis in Zulu, identify if the\
+  \ premise entails, contradicts, or is neutral towards the hypothesis. Please respond\
+  \ with exact 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_zul_prompt_3

lm_eval/tasks/afrixnli/translate/prompt_3/utils.py

+from lm_eval.utils import weighted_f1_score
+
+
+def doc_to_text(doc):
+    output = """You are an NLP assistant whose purpose is to solve Natural Language Inference (NLI) problems
+
+    Please identify whether the premise entails or contradicts the hypothesis in the following premise
+    and hypothesis. The answer should be exact entailment, contradiction, or neutral.
+
+    Premise: {premise}
+    Hypothesis: {hypothesis}
+
+    Is it entailment, contradiction, or neutral?"""
+
+    text = output.format(premise=doc["premise"], hypothesis=doc["hypothesis"])
+    return text
+
+
+def doc_to_target(doc):
+    replacements = {0: "entailment", 1: "neutral", 2: "contradiction"}
+    return replacements[doc["label"]]

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_amh.yaml

+# Generated by utils.py
+dataset_name: amh
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Amharic language.\nAnalyze the premise and hypothesis given in Amharic, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_amh_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_ewe.yaml

+# Generated by utils.py
+dataset_name: ewe
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Ewe language.\nAnalyze the premise and hypothesis given in Ewe, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_ewe_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_fra.yaml

+# Generated by utils.py
+dataset_name: fra
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the French language.\nAnalyze the premise and hypothesis given in French, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_fra_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_hau.yaml

+# Generated by utils.py
+dataset_name: hau
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Hausa language.\nAnalyze the premise and hypothesis given in Hausa, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_hau_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_ibo.yaml

+# Generated by utils.py
+dataset_name: ibo
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Igbo language.\nAnalyze the premise and hypothesis given in Igbo, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_ibo_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_kin.yaml

+# Generated by utils.py
+dataset_name: kin
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Kinyarwanda language.\nAnalyze the premise and hypothesis given in Kinyarwanda,\
+  \ and determine the relationship between them.\n Respond with one of the following\
+  \ options: 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_kin_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_lin.yaml

+# Generated by utils.py
+dataset_name: lin
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Lingala language.\nAnalyze the premise and hypothesis given in Lingala, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_lin_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_lug.yaml

+# Generated by utils.py
+dataset_name: lug
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Luganda language.\nAnalyze the premise and hypothesis given in Luganda, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_lug_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_orm.yaml

+# Generated by utils.py
+dataset_name: orm
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Oromo language.\nAnalyze the premise and hypothesis given in Oromo, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_orm_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_sna.yaml

+# Generated by utils.py
+dataset_name: sna
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the chiShona language.\nAnalyze the premise and hypothesis given in chiShona,\
+  \ and determine the relationship between them.\n Respond with one of the following\
+  \ options: 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}}\
+  \ \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_sna_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_sot.yaml

+# Generated by utils.py
+dataset_name: sot
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Sesotho language.\nAnalyze the premise and hypothesis given in Sesotho, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_sot_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_swa.yaml

+# Generated by utils.py
+dataset_name: swa
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Swahili language.\nAnalyze the premise and hypothesis given in Swahili, and\
+  \ determine the relationship between them.\n Respond with one of the following options:\
+  \ 'entailment', 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis:\
+  \ {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_swa_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_twi.yaml

+# Generated by utils.py
+dataset_name: twi
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Twi language.\nAnalyze the premise and hypothesis given in Twi, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_twi_prompt_4

lm_eval/tasks/afrixnli/translate/prompt_4/afrixnli_translate_wol.yaml

+# Generated by utils.py
+dataset_name: wol
+doc_to_text: "You are an expert in Natural Language Inference (NLI) specializing in\
+  \ the Wolof language.\nAnalyze the premise and hypothesis given in Wolof, and determine\
+  \ the relationship between them.\n Respond with one of the following options: 'entailment',\
+  \ 'contradiction', or 'neutral'. \n\nPremise: {{premise}} \nHypothesis: {{hypothesis}}"
+include: afrixnli_translate_yaml
+task: afrixnli_translate_wol_prompt_4