Elo-based RAG Agent evaluator
RAGElo1 is a streamlined toolkit for evaluating Retrieval Augmented Generation (RAG)-powered Large Language Models (LLMs) question answering agents using the Elo rating system.
While it has become easier to prototype and incorporate generative LLMs in production, evaluation is still the most challenging part of the solution. Comparing different outputs from multiple prompt and pipeline variations to a "gold standard" is not easy. Still, we can ask a powerful LLM to judge between pairs of answers and a set of questions.
This led us to develop a simple tool for tournament-style Elo ranking of LLM outputs. By comparing answers from different RAG pipelines and prompts over multiple questions, RAGElo computes a ranking of the different settings, providing a good overview of what works (and what doesn't).
For using RAGElo as a Python library, install it using pip:
pip install ragelo
If you want to (also) use RAGElo as a standalone CLI app, use the [cli]
tag:
pip install ragelo[cli]
To use RAGElo as a library, all you need to do is import RAGElo, initialize an Evaluator
and call either evaluate()
for evaluating a retrieved document or an LLM answer, or batch_evaluate()
to evaluate multiple responses at once. For example, using the RDNAM
retrieval evaluator from the Thomas et al. (2023) paper on using GPT-4 for annotating retrieval results:
from ragelo import get_retrieval_evaluator
evaluator = get_retrieval_evaluator("RDNAM", llm_provider="openai")
raw_answer, processed_answer = evaluator.evaluate(query="What is the capital of France?", document='Lyon is the second largest city in France.')
print(processed_answer)
# Output: 1
print(raw_answer)
# Output: '"O": 1\n}'
For a more complete example, we can evaluate with a custom prompt, and inject metadata into our evaluation prompt:
from ragelo import get_retrieval_evaluator
prompt = """You are a helpful assistant for evaluating the relevance of a retrieved document to a user query.
You should pay extra attention to how **recent** a document is. A document older than 5 years is considered outdated.
The answer should be evaluated according tot its recency, truthfulness, and relevance to the user query.
User query: {q}
Retrieved document: {d}
The document has a date of {document_date}.
Today is {today_date}.
WRITE YOUR ANSWER ON A SINGLE LINE AS A JSON OBJECT WITH THE FOLLOWING KEYS:
- "relevance": 0 if the document is irrelevant, 1 if it is relevant.
- "recency": 0 if the document is outdated, 1 if it is recent.
- "truthfulness": 0 if the document is false, 1 if it is true.
- "reasoning": A short explanation of why you think the document is relevant or irrelevant.
"""
evaluator = get_retrieval_evaluator(
"custom_prompt", # name of the retrieval evaluator
llm_provider="openai", # Which LLM provider to use
prompt=prompt, # your custom prompt
query_placeholder="q", # the placeholder for the query in the prompt
document_placeholder="d", # the placeholder for the document in the prompt
answer_format="multi_field_json", # The format of the answer. In this case, a JSON object with multiple fields
scoring_keys=["relevance", "recency", "truthfulness", "reasoning"], # Which keys to extract from the answer
)
raw_answer, answer = evaluator.evaluate(
query="What is the capital of Brazil?", # The user query
document="Rio de Janeiro is the capital of Brazil.", # The retrieved document
query_metadata={"today_date": "08-04-2024"}, # Some metadata for the query
doc_metadata={"document_date": "04-03-1950"}, # Some metadata for the document
)
answer
{'relevance': 0,
'recency': 0,
'truthfulness': 0,
'reasoning': 'The document is outdated and incorrect. Rio de Janeiro was the capital of Brazil until 1960 when it was changed to Brasília.'}
Note that, in this example, we passed to the evaluate
method two dictionaries with metadata for the query and the document. This metadata is injected into the prompt by matching their keys into the placeholders in the prompt.
After installing RAGElo as a CLI app, you can run it with the following command:
ragelo run-all queries.csv documents.csv answers.csv --verbose
---------- Agent Scores by Elo ranking ----------
agent1 : 1026.7
agent2 : 973.3
When running as a CLI, RAGElo expects the input files as CSV files. Specifically, it needs a csv file with the user queries, one with the documents retrieved by the retrieval system and one of the answers each agent produced. Here are some examples of the expected format:
queries.csv
:
qid,query
0, What is the capital of Brazil?
1, What is the capital of France?
documents.csv
:
qid,did,document_text
0,0, Brasília is the capital of Brazil.
0,1, Rio de Janeiro used to be the capital of Brazil.
1,2, Paris is the capital of France.
1,3, Lyon is the second largest city in France.
answers.csv
:
qid,agent,answer
0, agent1,"Brasília is the capital of Brazil, according to [0]."
0, agent2,"According to [1], Rio de Janeiro used to be the capital of Brazil, until the 60s."
1, agent1,"Paris is the capital of France, according to [2]."
1, agent2,"According to [3], Lyon is the second largest city in France. Meanwhile, Paris is its capital [2]."
While RAGElo can be used as either an end-to-end tool or by calling individual CLI components.
The retrieval-evaluator
tool annotates retrieved documents based on their relevance to the user query. This is done regardless of the answers provided by any Agent. As an example, for calling the Reasoner
retrieval evaluator (reasoner only outputs the reasoning why a document is relevant or not) we can use:
ragelo retrieval-evaluator reasoner queries.csv documents.csv output.csv
The output file changes according to the evaluator used. In general it will have one row per document evaluator, with the query_id, document_id, the raw LLM answer and the parsed answer. An example of the output for the reasoner is found here: tests/data/reasonings.csv.
The answers-annotator
tool annotates the answers generated by the Agents, taking the quality of the documents retrieved by the retrieval pipeline. By default, it uses the PairwiseWithReasoning
annotator, which generates k
random pairs of answers for each query and chooses the best answer based on the relevant documents cited in the answer. It relies on the reasonings.csv
file generated by the documents-annotator
:
ragelo answer-evaluator pairwise-reasoning queries.csv answers.csv reasonings.csv pairwise_answers_evaluations.csv
The pairwise_answers_evaluations.csv
file is a CSV file with both the raw answer and the parsed result for each pair of "games" between two agents. An output file example is provided at tests/data/pairwise_answers_evaluations.csv
Finally, the agents-ranker
tool ranks the agents by simulating an Elo tournament where the output of each game is given by the answers from the answers-annotator
:
ragelo agents-ranker eloranker pairwise_answers_evaluations.csv agents_ranking.csv
The output of this step is written to the output file agents_ranking.csv
with columns agent and score: tests/data/agents_ranking.csv.
To install the development dependencies, download the repository and run the following:
git clone https://github.com/zeta-alpha/ragelo && cd ragelo
pip install -e '.[dev]'
This will install the requirement dependencies in an editable mode (i.e., any changes to the code don't need to be rebuilt.)
For building a new version, use the build
command:
python -m build
- Add CI/CD for publishing
- Add full documentation of all implemented Evaluators
- Add option to few-shot examples (Undocumented, yet)
- Testing!
- Publish on PyPi
- Add more document evaluators (Microsoft)
- Split Elo evaluator
- Install as standalone CLI
Footnotes
-
The RAGElo logo was created using Dall-E 3 and GPT-4 with the following prompt: "Vector logo design for a toolkit named 'RAGElo'. The logo should have bold, modern typography with emphasis on 'RAG' in a contrasting color. Include a minimalist icon symbolizing retrieval or ranking." ↩