Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add ragbench faithfulness cards #1598

Open
wants to merge 9 commits into
base: main
Choose a base branch
from
Open

add ragbench faithfulness cards #1598

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
3104bed
allow multi reference in hhem metric
lilacheden Feb 11, 2025
8144ebd
add ragbench faithfulness cards
lilacheden Feb 11, 2025
7e531ef
add "mistral-large-instruct" to provider
lilacheden Feb 11, 2025
16ebe42
add "mistral-large-instruct" classification engines
lilacheden Feb 11, 2025
56347a8
add rag judges that use mistral-large-instruct
lilacheden Feb 11, 2025
d6105d6
Merge branch 'main' into ragbench
lilacheden Feb 13, 2025
0abc518
fix hhem multi reference
lilacheden Feb 13, 2025
1bd0a0a
Revert "fix hhem multi reference"
lilacheden Feb 13, 2025
ff9fa11
fix hhem multi reference
lilacheden Feb 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 41 additions & 0 deletions prepare/cards/ragbench_faithfulness.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
from unitxt import add_to_catalog
from unitxt.blocks import (
LoadHF,
TaskCard,
)
from unitxt.operators import Copy, ExecuteExpression
from unitxt.templates import NullTemplate

for subset in [
"covidqa",
"cuad",
"delucionqa",
"emanual",
"expertqa",
"finqa",
"hagrid",
"hotpotqa",
"msmarco",
"pubmedqa",
"tatqa",
"techqa",
]:
card = TaskCard(
loader=LoadHF(
path="rungalileo/ragbench",
name=subset,
split="test"
),
preprocess_steps=[
Copy(field="response", to_field="answer"),
Copy(field="documents", to_field="contexts"),
ExecuteExpression(expression="int(adherence_score)", to_field="number_val"),
ExecuteExpression(expression="['yes' if adherence_score else 'no']", to_field="is_faithful"),
],
task="tasks.rag_eval.faithfulness.binary",
templates={"default": NullTemplate()},
)

add_to_catalog(
card, f"cards.rag_eval.faithfulness.ragbench.{subset}", overwrite=True
)
1 change: 1 addition & 0 deletions prepare/engines/classification/classification_engines.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
)

model_names_to_provider = {
"mistral-large-instruct": ["watsonx", "rits"],
"llama-3-3-70b-instruct": ["watsonx", "rits"],
"llama-3-1-70b-instruct": ["watsonx", "rits"],
"gpt-4o": ["open-ai"],
Expand Down
2 changes: 2 additions & 0 deletions prepare/metrics/llm_as_judge/rag_judge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -102,6 +102,8 @@ def get_prediction_field(metric_type):
"llama_3_3_70b_instruct_watsonx": "engines.classification.llama_3_3_70b_instruct_watsonx",
"llama_3_3_70b_instruct_rits": "engines.classification.llama_3_3_70b_instruct_rits",
"gpt_4o_azure": "engines.classification.gpt_4o_2024_08_06_azure_openai",
"mistral_large_instruct_watsonx": "engines.classification.mistral_large_watsonx",
"mistral_large_instruct_rits": "engines.classification.mistral_large_instruct_2407_rits",
generic_engine_label: GenericInferenceEngine(),
}

Expand Down
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/covidqa.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "covidqa",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/cuad.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "cuad",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/delucionqa.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "delucionqa",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/emanual.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "emanual",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/expertqa.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "expertqa",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/finqa.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "finqa",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/hagrid.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "hagrid",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
37 changes: 37 additions & 0 deletions src/unitxt/catalog/cards/rag_eval/faithfulness/ragbench/hotpotqa.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
{
"__type__": "task_card",
"loader": {
"__type__": "load_hf",
"path": "rungalileo/ragbench",
"name": "hotpotqa",
"split": "test"
},
"preprocess_steps": [
{
"__type__": "copy",
"field": "response",
"to_field": "answer"
},
{
"__type__": "copy",
"field": "documents",
"to_field": "contexts"
},
{
"__type__": "execute_expression",
"expression": "int(adherence_score)",
"to_field": "number_val"
},
{
"__type__": "execute_expression",
"expression": "['yes' if adherence_score else 'no']",
"to_field": "is_faithful"
}
],
"task": "tasks.rag_eval.faithfulness.binary",
"templates": {
"default": {
"__type__": "null_template"
}
}
}
Loading
Loading