We lifted 6 functions out of this paper's own repositories and ran 5 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| project-miracl/nomiracl | canonical | 5 of 6 |
| Function | Status | Where it lives |
|---|---|---|
| count_word | Ran | project-miracl/nomiracl/nomiracl/util.py code served (permissive licence) · get_code("dc273dd33d8e4d97") |
| load_corpus | Ran | project-miracl/nomiracl/nomiracl/dataset/util.py code served (permissive licence) · get_code("7ef73ee895d32f5b") |
| load_qrels | Ran | project-miracl/nomiracl/nomiracl/dataset/util.py code served (permissive licence) · get_code("d5328a85425a2266") |
| load_queries | Ran | project-miracl/nomiracl/nomiracl/dataset/util.py code served (permissive licence) · get_code("3f19e75689dbe916") |
| load_results_as_jsonl | Ran | project-miracl/nomiracl/nomiracl/util.py code served (permissive licence) · get_code("25236e0fcb34df87") |
| check_package_availability | Not yet run | project-miracl/nomiracl/nomiracl/util.py code served (permissive licence) · get_code("933fa94aeb8df31d") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Retrieval-Augmented Generation (RAG) grounds Large Language Model (LLM) output by leveraging external knowledge sources to reduce factual hallucinations. However, prior work lacks a comprehensive evaluation of different language families, making it challenging to evaluate LLM robustness against errors in external retrieved knowledge. To overcome this, we establish NoMIRACL, a human-annotated dataset for evaluating LLM robustness in RAG across 18 typologically diverse languages. NoMIRACL includes both a non-relevant and a relevant subset. Queries in the non-relevant subset contain passages judged as non-relevant, whereas queries in the relevant subset include at least a single judged relevant passage. We measure relevance assessment using: (i) hallucination rate, measuring model tendency to hallucinate, when the answer is not present in passages in the non-relevant subset, and (ii) error rate, measuring model inaccuracy to recognize relevant passages in the relevant subset.In our work, we observe that most models struggle to balance the two capacities. Models such as LLAMA-2 and Orca-2 achieve over 88% hallucination rate on the non-relevant subset. Mistral and LLAMA-3 hallucinate less but can achieve up to a 74.9% error rate on the relevant subset. Overall, GPT-4 is observed to provide the best tradeoff on both subsets, highlighting future work necessary to improve LLM robustness. NoMIRACL dataset and evaluation code are available at: https://github.com/project-miracl/nomiracl.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2312.11361")
get_code_for_paper("2312.11361")
have("2312.11361")
Connect an agent — have() is free.