Datasets:

ArsSocratica
/

egora-benchmarks

Tasks:

Text Generation

Size:

ArXiv:

Tags:

rotation-retention-law

knowledge-retention

License:

Dataset card Files Files and versions

egora-benchmarks

20.3 MB

2 contributors

History: 5 commits

ArsSocratica

Update README: only Llama, Mistral, Phi (paper models)

f70efd7 about 2 months ago

cross-modal
Remove LLaVA and Qwen2-VL data (not in paper) about 2 months ago
llama-3.1-8b
Upload EgoRA benchmark results: Llama 1B/3B/8B (Alpaca+Medical), cross-modal (Mistral-7B, Phi-3 Mini, LLaVA-7B, Qwen2-VL-7B), rotation geometry, threshold analysis about 2 months ago
llama-3.2-1b
Upload EgoRA benchmark results: Llama 1B/3B/8B (Alpaca+Medical), cross-modal (Mistral-7B, Phi-3 Mini, LLaVA-7B, Qwen2-VL-7B), rotation geometry, threshold analysis about 2 months ago
llama-3.2-3b
Upload EgoRA benchmark results: Llama 1B/3B/8B (Alpaca+Medical), cross-modal (Mistral-7B, Phi-3 Mini, LLaVA-7B, Qwen2-VL-7B), rotation geometry, threshold analysis about 2 months ago
threshold-analysis
Upload EgoRA benchmark results: Llama 1B/3B/8B (Alpaca+Medical), cross-modal (Mistral-7B, Phi-3 Mini, LLaVA-7B, Qwen2-VL-7B), rotation geometry, threshold analysis about 2 months ago
.gitattributes

2.5 kB
initial commit about 2 months ago
README.md

2.79 kB
Update README: only Llama, Mistral, Phi (paper models) about 2 months ago