Kyle Ng
kylelovesllms
AI & ML interests
None yet
Recent Activity
updated a collection 10 days ago
translation-syntax-datasets-03_grammar_v3 updated a dataset 10 days ago
kylelovesllms/grammar-v3-frames-rb published a dataset 10 days ago
kylelovesllms/grammar-v3-frames-rbOrganizations
Based off of `tatsu-lab/alpaca` and `yahma/alpaca-cleaned` and translated datasets
Qwen Models fine tuned using LoRA Adapters on Alpaca dataset in English, German, and Russian to test cross-linguistic generalization of fine tuning
-
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-en-lora
Text Generation • Updated • 20 -
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-ru-lora-merged
0.5B • Updated • 9 -
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-en-lora-merged
Text Generation • 0.5B • Updated • 16 -
kylelovesllms/Qwen2.5-1.5B-Instruct-caps-ru-lora-merged
2B • Updated • 9
Translated versions of `sorry-bench/sorry-bench-202406` to fine-tune on jail broken models
Jailbroken Qwen fine-tuned on sorry-bench/sorry-bench-202406 (English, Chinese, French) to test cross-linguistic generalization of refusal tasks
-
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-merged
0.5B • Updated • 4 -
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-fr-merged
0.5B • Updated • 8 -
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-zh-cn-merged
0.5B • Updated • 8 -
kylelovesllms/Qwen2.5-1.5B-Instruct-refusals-lora-zh-cn-merged
2B • Updated • 4
Translated versions of `sorry-bench/sorry-bench-202406` to fine-tune on jail broken models
Based off of `tatsu-lab/alpaca` and `yahma/alpaca-cleaned` and translated datasets
Jailbroken Qwen fine-tuned on sorry-bench/sorry-bench-202406 (English, Chinese, French) to test cross-linguistic generalization of refusal tasks
-
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-merged
0.5B • Updated • 4 -
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-fr-merged
0.5B • Updated • 8 -
kylelovesllms/Qwen2.5-0.5B-Instruct-Jailbroken-refusals-lora-zh-cn-merged
0.5B • Updated • 8 -
kylelovesllms/Qwen2.5-1.5B-Instruct-refusals-lora-zh-cn-merged
2B • Updated • 4
Qwen Models fine tuned using LoRA Adapters on Alpaca dataset in English, German, and Russian to test cross-linguistic generalization of fine tuning
-
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-en-lora
Text Generation • Updated • 20 -
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-ru-lora-merged
0.5B • Updated • 9 -
kylelovesllms/Qwen2.5-0.5B-Instruct-caps-en-lora-merged
Text Generation • 0.5B • Updated • 16 -
kylelovesllms/Qwen2.5-1.5B-Instruct-caps-ru-lora-merged
2B • Updated • 9