Archangel is a suite of human feedback-aligned LLMs, released as part of the Human-Aware Loss Functions (HALOs) project by Ethayarajh et al. (2024).
ContextualAI
company
Verified
AI & ML interests
Enterprise-grade retrieval augmented language models
models
81
ContextualAI/tiny-random-MistralForCausalLM
Text Generation
•
Updated
•
452
ContextualAI/zephyr_sft_dpo
Text Generation
•
Updated
•
2
ContextualAI/zephyr_sft_kto_unary
Text Generation
•
Updated
•
2
ContextualAI/zephyr_sft_kto
Text Generation
•
Updated
•
2
•
1
ContextualAI/Contextual_KTO_Mistral_PairRM
Text Generation
•
Updated
•
781
•
25
ContextualAI/archangel_kto_pythia6-9b
Text Generation
•
Updated
ContextualAI/archangel_kto_pythia2-8b
Text Generation
•
Updated
ContextualAI/archangel_kto_pythia1-4b
Text Generation
•
Updated
•
1
ContextualAI/archangel_ppo_pythia2-8b
Text Generation
•
Updated
•
1
ContextualAI/archangel_ppo_pythia1-4b
Text Generation
•
Updated
•
4
datasets
10
ContextualAI/ultrabin_clean_max_chosen_min_rejected_rationalized
Viewer
•
Updated
•
60.9k
•
5
ContextualAI/ultrabin_clean_max_chosen_rand_rejected_rationalized
Viewer
•
Updated
•
60.9k
ContextualAI/ultrabin_clean_max_chosen_min_rejected_rationalized_helpfulness
Viewer
•
Updated
•
60.9k
•
1.66k
ContextualAI/ultrabin_clean_max_chosen_min_rejected_rationalized_truthfulness
Viewer
•
Updated
•
60.9k
•
1.62k
ContextualAI/ultrabin_clean_max_chosen_min_rejected_rationalized_honesty
Viewer
•
Updated
•
60.9k
•
1.58k
ContextualAI/ultrabin_clean_max_chosen_min_rejected_rationalized_instruction_following
Viewer
•
Updated
•
60.9k
•
1.61k
ContextualAI/ultrabin_clean_max_chosen_min_rejected
Viewer
•
Updated
•
60.9k
•
2.29k
ContextualAI/ultrabin_clean_max_chosen_rand_rejected
Viewer
•
Updated
•
60.9k
•
2.75k
ContextualAI/ultrafeedback_binarized_1percent
Viewer
•
Updated
•
200
•
346
ContextualAI/tiny_train_test_ultrafeedback_binarized
Viewer
•
Updated
•
20
•
4