# Causal Question Answering on CausalProbe-H Source: [Causal Question Answering on CausalProbe-H]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:41:32.135Z Evaluation of model performance on the CausalProbe-H dataset for Causal Question Answering tasks, typically measured by Exact Match (EM). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Causal Question Answering | | Dataset (including any stated split) | CausalProbe-H | | Metric | EM | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 32 of 32 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [EM](): Higher is better; unit: SCALAR. ## Recorded results | Method | EM | Result date | Report paper | | --- | --- | --- | --- | | G²-Reasoner | 0.701 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.701 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.7 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.696 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.693 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.692 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.69 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.688 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.687 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.685 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.678 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.678 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.671 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.67 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.667 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.662 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.658 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.658 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.655 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.655 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.652 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.652 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.644 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.644 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.582 | Not recorded | 68fa79b31e209131cea7ef6d | | G²-Reasoner | 0.582 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.575 | Not recorded | 68fa79b31e209131cea7ef6d | | RAG | 0.575 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.573 | Not recorded | 68fa79b31e209131cea7ef6d | | CoT | 0.573 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.565 | Not recorded | 68fa79b31e209131cea7ef6d | | Vanilla | 0.565 | Not recorded | 68fa79b31e209131cea7ef6d | ### G²-Reasoner EM: 0.701. Result ID: 698b59957a967eddb76ddc12. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Claude 3 opus","Temperature":"1.0"}. This record does not assert independent reproduction. ### CoT EM: 0.701. Result ID: 698c535b37894ef7d9cd8c44. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Claude 3"}. This record does not assert independent reproduction. ### CoT EM: 0.7. Result ID: 698b59957a967eddb76ddc10. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Claude 3 opus","Temperature":"1.0"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.696. Result ID: 698c535b37894ef7d9cd8c46. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Claude 3"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.693. Result ID: 698c535b37894ef7d9cd8c42. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT 3.5"}. This record does not assert independent reproduction. ### Vanilla EM: 0.692. Result ID: 698c535b37894ef7d9cd8c43. [Reporting paper]() Recorded evaluation configuration: {"Model":"Claude 3"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.69. Result ID: 698b59957a967eddb76ddc0e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"GPT 3.5 turbo","Temperature":"1.0"}. This record does not assert independent reproduction. ### Vanilla EM: 0.688. Result ID: 698b59957a967eddb76ddc0f. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Claude 3 opus","Temperature":"1.0"}. This record does not assert independent reproduction. ### RAG EM: 0.687. Result ID: 698b59957a967eddb76ddc11. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Claude 3 opus","Temperature":"1.0"}. This record does not assert independent reproduction. ### RAG EM: 0.685. Result ID: 698c535b37894ef7d9cd8c45. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Claude 3"}. This record does not assert independent reproduction. ### RAG EM: 0.678. Result ID: 698b59957a967eddb76ddc0d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"GPT 3.5 turbo","Temperature":"1.0"}. This record does not assert independent reproduction. ### RAG EM: 0.678. Result ID: 698c535b37894ef7d9cd8c41. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT 3.5"}. This record does not assert independent reproduction. ### Vanilla EM: 0.671. Result ID: 698c535b37894ef7d9cd8c3f. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT 3.5"}. This record does not assert independent reproduction. ### Vanilla EM: 0.67. Result ID: 698b59957a967eddb76ddc0b. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"GPT 3.5 turbo","Temperature":"1.0"}. This record does not assert independent reproduction. ### CoT EM: 0.667. Result ID: 698c535b37894ef7d9cd8c40. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT 3.5"}. This record does not assert independent reproduction. ### CoT EM: 0.662. Result ID: 698b59957a967eddb76ddc0c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT 3.5 turbo","Temperature":"1.0"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.658. Result ID: 698b59957a967eddb76ddc0a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA 3"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.658. Result ID: 698c535b37894ef7d9cd8c3e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"LLaMA 3 8B"}. This record does not assert independent reproduction. ### RAG EM: 0.655. Result ID: 698b59957a967eddb76ddc09. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA 3"}. This record does not assert independent reproduction. ### RAG EM: 0.655. Result ID: 698c535b37894ef7d9cd8c3d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"LLaMA 3 8B"}. This record does not assert independent reproduction. ### Vanilla EM: 0.652. Result ID: 698b59957a967eddb76ddc07. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaMA 3"}. This record does not assert independent reproduction. ### Vanilla EM: 0.652. Result ID: 698c535b37894ef7d9cd8c3b. [Reporting paper]() Recorded evaluation configuration: {"Model":"LLaMA 3 8B"}. This record does not assert independent reproduction. ### CoT EM: 0.644. Result ID: 698b59957a967eddb76ddc08. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA 3"}. This record does not assert independent reproduction. ### CoT EM: 0.644. Result ID: 698c535b37894ef7d9cd8c3c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LLaMA 3 8B"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.582. Result ID: 698b59957a967eddb76ddc06. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA 2"}. This record does not assert independent reproduction. ### G²-Reasoner EM: 0.582. Result ID: 698c535b37894ef7d9cd8c3a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"LLaMA 2 7B"}. This record does not assert independent reproduction. ### RAG EM: 0.575. Result ID: 698b59957a967eddb76ddc05. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA 2"}. This record does not assert independent reproduction. ### RAG EM: 0.575. Result ID: 698c535b37894ef7d9cd8c39. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"LLaMA 2 7B"}. This record does not assert independent reproduction. ### CoT EM: 0.573. Result ID: 698b59957a967eddb76ddc04. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA 2"}. This record does not assert independent reproduction. ### CoT EM: 0.573. Result ID: 698c535b37894ef7d9cd8c38. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LLaMA 2 7B"}. This record does not assert independent reproduction. ### Vanilla EM: 0.565. Result ID: 698b59957a967eddb76ddc03. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"LLaMA 2"}. This record does not assert independent reproduction. ### Vanilla EM: 0.565. Result ID: 698c535b37894ef7d9cd8c37. [Reporting paper]() Recorded evaluation configuration: {"Model":"LLaMA 2 7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.