Model Card for NordicAlpaca-Finnish-V1-7B

This is a merge of mpasila/NordicAlpaca-Finnish-V1-7B-LoRA.

Dataset used with the LoRA is pinzhenchen/alpaca-cleaned-fi.

Base model used: HPLT/gpt-7b-nordic-prerelease

It uses Alpaca format but with a translated instruction at the start:

{
    "instruction,output": "Alla on ohje, jossa kuvataan tehtävä. Kirjoita vastaus, joka täyttää pyynnön asianmukaisesti.\n\n### Instruction:\n%instruction%\n\n### Response:\n%output%",
    "instruction,input,output": "Alla on ohje, jossa kuvataan tehtävä ja joka on yhdistetty kontekstia lisäävään syötteeseen. Kirjoita vastaus, joka täyttää pyynnön asianmukaisesti.\n\n### Instruction:\n%instruction%\n\n### Input:\n%input%\n\n### Response:\n%output%"
}

Merged using this Colab notebook. It might not be the best way to merge a quantized LoRA on to a float16 model but I just wanted to quickly do something. You can try merging it better if you want.

Evaluation

Model	Size	Type	FIN-bench (score)
mpasila/NordicAlpaca-Finnish-V1-7B	7B	Instruct	0.3891
mpasila/Finnish-Alpaca-Small-7B	7B	Instruct	0.3586
mpasila/Finnish-Alpaca-Tiny-V2-7B	7B	Instruct	0.4654
mpasila/Alpacazord-Viking-7B	7B	Instruct	0.4123
mpasila/Finnish-Viking-Alpaca-V1-7B	7B	Instruct	0.3943
Finnish-NLP/llama-7b-finnish-instruct-v0.1	7B	Instruct	0.4365
Finnish-NLP/llama-7b-finnish-instruct-v0.2	7B	Instruct	0.3993
Finnish-NLP/llama-7b-finnish	7B	Base	0.2350
LumiOpen/Viking-7B (1000B)	7B	Base	0.3721
HPLT/gpt-7b-nordic-prerelease	7B	Base	0.3169

Source

FIN-bench scores:

Task	Version	Metric	Value		Stderr
bigbench_analogies	0	multiple_choice_grade	0.5615	±	0.0437
bigbench_arithmetic_1_digit_addition	0	multiple_choice_grade	0.5300	±	0.0502
bigbench_arithmetic_1_digit_division	0	multiple_choice_grade	0.8261	±	0.0808
bigbench_arithmetic_1_digit_multiplication	0	multiple_choice_grade	0.3700	±	0.0485
bigbench_arithmetic_1_digit_subtraction	0	multiple_choice_grade	0.6000	±	0.0492
bigbench_arithmetic_2_digit_addition	0	multiple_choice_grade	0.0300	±	0.0171
bigbench_arithmetic_2_digit_division	0	multiple_choice_grade	0.4700	±	0.0502
bigbench_arithmetic_2_digit_multiplication	0	multiple_choice_grade	0.1000	±	0.0302
bigbench_arithmetic_2_digit_subtraction	0	multiple_choice_grade	0.1800	±	0.0386
bigbench_arithmetic_3_digit_addition	0	multiple_choice_grade	0.3000	±	0.0461
bigbench_arithmetic_3_digit_division	0	multiple_choice_grade	0.2000	±	0.0402
bigbench_arithmetic_3_digit_multiplication	0	multiple_choice_grade	0.1800	±	0.0386
bigbench_arithmetic_3_digit_subtraction	0	multiple_choice_grade	0.2500	±	0.0435
bigbench_arithmetic_4_digit_addition	0	multiple_choice_grade	0.4200	±	0.0496
bigbench_arithmetic_4_digit_division	0	multiple_choice_grade	0.2400	±	0.0429
bigbench_arithmetic_4_digit_multiplication	0	multiple_choice_grade	0.2200	±	0.0416
bigbench_arithmetic_4_digit_subtraction	0	multiple_choice_grade	0.4600	±	0.0501
bigbench_arithmetic_5_digit_addition	0	multiple_choice_grade	0.5400	±	0.0501
bigbench_arithmetic_5_digit_division	0	multiple_choice_grade	0.1100	±	0.0314
bigbench_arithmetic_5_digit_multiplication	0	multiple_choice_grade	0.2400	±	0.0429
bigbench_arithmetic_5_digit_subtraction	0	multiple_choice_grade	0.5200	±	0.0502
bigbench_cause_and_effect_one_sentence	0	multiple_choice_grade	0.6471	±	0.0676
bigbench_cause_and_effect_one_sentence_no_prompt	0	multiple_choice_grade	0.8039	±	0.0561
bigbench_cause_and_effect_two_sentences	0	multiple_choice_grade	0.3529	±	0.0676
bigbench_emotions	0	multiple_choice_grade	0.2938	±	0.0361
bigbench_empirical_judgments	0	multiple_choice_grade	0.3333	±	0.0476
bigbench_general_knowledge	0	multiple_choice_grade	0.2857	±	0.0544
bigbench_hhh_alignment_harmless	0	multiple_choice_grade	0.3448	±	0.0630
bigbench_hhh_alignment_helpful	0	multiple_choice_grade	0.3220	±	0.0614
bigbench_hhh_alignment_honest	0	multiple_choice_grade	0.3729	±	0.0635
bigbench_hhh_alignment_other	0	multiple_choice_grade	0.5581	±	0.0766
bigbench_intent_recognition	0	multiple_choice_grade	0.1777	±	0.0145
bigbench_misconceptions	0	multiple_choice_grade	0.5373	±	0.0432
bigbench_paraphrase	0	multiple_choice_grade	0.4750	±	0.0354
bigbench_sentence_ambiguity	0	multiple_choice_grade	0.4333	±	0.0645
bigbench_similarities_abstraction	0	multiple_choice_grade	0.7237	±	0.0516

Framework versions

• PEFT 0.8.2