7 research outputs found
Recommended from our members
History in the literary imagination: the telling of Nongqawuse and the Xhosa Cattle-Killing in South African literature and culture (1891-1937)
No full textThis thesis takes as its subject the millenarian movement of 1856–7, commonly known as the Xhosa Cattle-Killing. My project examines a range of literary representations of this seminal moment in South African history: novels, plays, and short stories in English or English translation. The period under consideration encompasses the earliest literary responses to the Cattle-Killing and includes critical historical-political moments such as: the incorporation of the last independent black territory into the Cape Colony, the creation of the Union of South Africa, the passing of the Land Act, the enfranchisement of white women and the enactment of Hertzog’s ‘native bills’. The project consists of close, contextual readings, and the approach is cross-cultural and interdisciplinary.
In this dissertation I examine the meaning that has accrued to the Cattle-Killing, and the role that literary accounts have played in interpreting and defining this pivotal event in the historical consciousness of their sometimes considerable audiences. In some cases, these creative works have anticipated trends in formal historiography and suggested new ways to interrogate the evidence. But the accounts do more than creatively reconstruct the past. They are also implicated in their respective presents and use the Cattle-Killing to ‘write out’ contemporaneous concerns: be it female emancipation, ‘native education’ or Black Nationalism. The various manifestations of the Cattle-Killing story chart not only the shifting ‘truth’ of the event but also the ways in which it has been made relevant and useable for different communities at various points in South Africa’s history. To read these accounts of the Cattle-Killing, I argue, is to ‘read’ the history of this period.
While taking as its subject an event from 150 years ago, and literary responses from shortly after, my project contributes to wider, on-going conversations relating to history as a field of argument and literature as a social and historical force. A related aim is to contribute to the revaluation of early South African literature, which has been neglected or homogenized in recent years. My dissertation seeks to recuperate and complicate by representing a variety of subject positions and resuscitating voices discarded or forgotten
Humanity's Last Exam
No full textBenchmarks are important tools for tracking the rapid advancements in large
language model (LLM) capabilities. However, benchmarks are not keeping pace in
difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like
MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In
response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at
the frontier of human knowledge, designed to be the final closed-ended academic
benchmark of its kind with broad subject coverage. HLE consists of 2,700
questions across dozens of subjects, including mathematics, humanities, and the
natural sciences. HLE is developed globally by subject-matter experts and
consists of multiple-choice and short-answer questions suitable for automated
grading. Each question has a known solution that is unambiguous and easily
verifiable, but cannot be quickly answered via internet retrieval.
State-of-the-art LLMs demonstrate low accuracy and calibration on HLE,
highlighting a significant gap between current LLM capabilities and the expert
human frontier on closed-ended academic questions. To inform research and
policymaking upon a clear understanding of model capabilities, we publicly
release HLE at https://lastexam.ai
Humanity's Last Exam
No full textBenchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai
Humanity's Last Exam
No full textBenchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai
Recommended from our members
Humanity's Last Exam
No full textBenchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 2,500 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai
Humanity's Last Exam
No full textBenchmarks are important tools for tracking the rapid advancements in large
language model (LLM) capabilities. However, benchmarks are not keeping pace in
difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like
MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In
response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at
the frontier of human knowledge, designed to be the final closed-ended academic
benchmark of its kind with broad subject coverage. HLE consists of 2,500
questions across dozens of subjects, including mathematics, humanities, and the
natural sciences. HLE is developed globally by subject-matter experts and
consists of multiple-choice and short-answer questions suitable for automated
grading. Each question has a known solution that is unambiguous and easily
verifiable, but cannot be quickly answered via internet retrieval.
State-of-the-art LLMs demonstrate low accuracy and calibration on HLE,
highlighting a significant gap between current LLM capabilities and the expert
human frontier on closed-ended academic questions. To inform research and
policymaking upon a clear understanding of model capabilities, we publicly
release HLE at https://lastexam.ai
A benchmark of expert-level academic questions to assess AI capabilities
No full textInternational audienceBenchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve more than 90% accuracy on popular benchmarks such as Measuring Massive Multitask Language Understanding1, limiting informed measurement of state-of-the-art LLM capabilities. Here, in response, we introduce Humanity’s Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be an expert-level closed-ended academic benchmark with broad subject coverage. HLE consists of 2,500 questions across dozens of subjects, including mathematics, humanities and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable but cannot be quickly answered by internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a marked gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai
