Information Retrieval on QUEST (test)
30.8P@1OrLog
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| OrLogBase Retriever=E5, Reasoning Strategy=OrLog, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 30.8 | 13.1 | 3.9 | 14.2 | 6.3 | 12.1 | 30.8 | 19.7 | 38.4 | |
| LLMBase Retriever=E5, Reasoning Strategy=LLM-as-reasoner, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 26.3 | 13.3 | 3.4 | 14.5 | 5.6 | 12.3 | 26.3 | 19.3 | 35.5 | |
| LLM*Base Retriever=E5, Reasoning Strategy=Logic-augmented LLM, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 24.1 | 12.7 | 3.1 | 13.8 | 5 | 11.8 | 24.1 | 18 | 33.3 | |
| OrLogBase Retriever=BM25, Reasoning Strategy=OrLog, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 22.1 | 8.9 | 2.6 | 9.5 | 4.4 | 8.2 | 22.1 | 13.4 | 26.7 | |
| LLM*Base Retriever=E5, Reasoning Strategy=Logic-augmented LLM, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 20.4 | 11.9 | 2.4 | 13.2 | 4.1 | 11.1 | 20.4 | 16.3 | 29.9 | |
| LLMBase Retriever=E5, Reasoning Strategy=LLM-as-reasoner, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 19.2 | 11.5 | 2.3 | 12.6 | 3.8 | 10.7 | 19.2 | 15.6 | 28.7 | |
| LLMBase Retriever=BM25, Reasoning Strategy=LLM-as-reasoner, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 18.5 | 8.6 | 2.3 | 9.3 | 3.8 | 8 | 18.5 | 12.5 | 24.4 | |
| E5Base Retriever=E5, Reasoning Strategy=Original Ranking2026.01 | 18.2 | 10.5 | 2.1 | 11.6 | 3.5 | 9.8 | 18.2 | 14.2 | 26.9 | |
| OrLogBase Retriever=E5, Reasoning Strategy=OrLog, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 16.1 | 10.4 | 1.9 | 11.4 | 3.1 | 9.7 | 16.1 | 13.6 | 26 | |
| LLM*Base Retriever=BM25, Reasoning Strategy=Logic-augmented LLM, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 15.8 | 8.3 | 1.9 | 8.8 | 3.1 | 7.7 | 15.8 | 11.4 | 21.9 | |
| LLMBase Retriever=BM25, Reasoning Strategy=LLM-as-reasoner, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 15.5 | 7.9 | 2 | 8.5 | 3.2 | 7.4 | 15.5 | 11 | 21.2 | |
| LLM*Base Retriever=BM25, Reasoning Strategy=Logic-augmented LLM, Knowledge Setting=Parametric+, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 14.4 | 7.6 | 1.9 | 8.3 | 3 | 7.1 | 14.4 | 10.6 | 20.3 | |
| OrLogBase Retriever=BM25, Reasoning Strategy=OrLog, Knowledge Setting=Parametric, Backbone Model=Llama-3.3-70B-Instruct2026.01 | 13.8 | 7.5 | 1.5 | 8 | 2.6 | 7 | 13.8 | 10.1 | 19.9 | |
| BM25Base Retriever=BM25, Reasoning Strategy=Original Ranking2026.01 | 11.6 | 6.7 | 1.4 | 7.3 | 2.2 | 6.2 | 11.6 | 8.9 | 17.5 |