--- license: apache-2.0 datasets: - totally-not-an-llm/everything-sharegptformat-morecleaned language: - en pipeline_tag: text-generation --- This is [OpenLLaMA 3B V2](https://huggingface.co/openlm-research/open_llama_3b_v2) finetuned on [EverythingLM Data(ShareGPT format more cleaned)](https://huggingface.co/datasets/totally-not-an-llm/everything-sharegptformat-morecleaned) for 1 epochs. Prompt template: ``` ### HUMAN: {prompt} ### RESPONSE: ``` GGML quants available [here](https://huggingface.co/TheBloke/Marx-3b-GGML).
All GGML quants available [here](https://huggingface.co/TheBloke/Marx-3b-GPTQ). Note: Don't expect this model to be good, I was just starting out to finetune. So don't roast me please! Benchmarks:
arc_challenge
acc0.38993174061433444
acc_stderr0.014252959848892884
acc_norm0.4308873720136519
acc_norm_stderr0.014471133392642475
hellaswag
acc0.5513841864170484
acc_stderr0.004963362085275556
acc_norm0.7257518422624976
acc_norm_stderr0.00445222854104355
hendrycksTest-abstract_algebra
acc0.23
acc_stderr0.04229525846816506
acc_norm0.23
acc_norm_stderr0.04229525846816506
hendrycksTest-anatomy
acc0.2962962962962963
acc_stderr0.03944624162501116
acc_norm0.2962962962962963
acc_norm_stderr0.03944624162501116
hendrycksTest-astronomy
acc0.32894736842105265
acc_stderr0.03823428969926603
acc_norm0.32894736842105265
acc_norm_stderr0.03823428969926603
hendrycksTest-business_ethics
acc0.3
acc_stderr0.046056618647183814
acc_norm0.3
acc_norm_stderr0.046056618647183814
hendrycksTest-clinical_knowledge
acc0.2641509433962264
acc_stderr0.027134291628741713
acc_norm0.2641509433962264
acc_norm_stderr0.027134291628741713
hendrycksTest-college_biology
acc0.2569444444444444
acc_stderr0.03653946969442099
acc_norm0.2569444444444444
acc_norm_stderr0.03653946969442099
hendrycksTest-college_chemistry
acc0.22
acc_stderr0.041633319989322695
acc_norm0.22
acc_norm_stderr0.041633319989322695
hendrycksTest-college_computer_science
acc0.26
acc_stderr0.0440844002276808
acc_norm0.26
acc_norm_stderr0.0440844002276808
hendrycksTest-college_mathematics
acc0.31
acc_stderr0.04648231987117316
acc_norm0.31
acc_norm_stderr0.04648231987117316
hendrycksTest-college_medicine
acc0.23121387283236994
acc_stderr0.032147373020294696
acc_norm0.23121387283236994
acc_norm_stderr0.032147373020294696
hendrycksTest-college_physics
acc0.27450980392156865
acc_stderr0.04440521906179327
acc_norm0.27450980392156865
acc_norm_stderr0.04440521906179327
hendrycksTest-computer_security
acc0.36
acc_stderr0.048241815132442176
acc_norm0.36
acc_norm_stderr0.048241815132442176
hendrycksTest-conceptual_physics
acc0.2765957446808511
acc_stderr0.029241883869628837
acc_norm0.2765957446808511
acc_norm_stderr0.029241883869628837
hendrycksTest-econometrics
acc0.2631578947368421
acc_stderr0.04142439719489363
acc_norm0.2631578947368421
acc_norm_stderr0.04142439719489363
hendrycksTest-electrical_engineering
acc0.20689655172413793
acc_stderr0.03375672449560554
acc_norm0.20689655172413793
acc_norm_stderr0.03375672449560554
hendrycksTest-elementary_mathematics
acc0.2698412698412698
acc_stderr0.022860838309232072
acc_norm0.2698412698412698
acc_norm_stderr0.022860838309232072
hendrycksTest-formal_logic
acc0.2619047619047619
acc_stderr0.039325376803928704
acc_norm0.2619047619047619
acc_norm_stderr0.039325376803928704
hendrycksTest-global_facts
acc0.35
acc_stderr0.047937248544110196
acc_norm0.35
acc_norm_stderr0.047937248544110196
hendrycksTest-high_school_biology
acc0.24193548387096775
acc_stderr0.0243625996930311
acc_norm0.24193548387096775
acc_norm_stderr0.0243625996930311
hendrycksTest-high_school_chemistry
acc0.28078817733990147
acc_stderr0.0316185633535861
acc_norm0.28078817733990147
acc_norm_stderr0.0316185633535861
hendrycksTest-high_school_computer_science
acc0.33
acc_stderr0.04725815626252605
acc_norm0.33
acc_norm_stderr0.04725815626252605
hendrycksTest-high_school_european_history
acc0.296969696969697
acc_stderr0.03567969772268048
acc_norm0.296969696969697
acc_norm_stderr0.03567969772268048
hendrycksTest-high_school_geography
acc0.2878787878787879
acc_stderr0.03225883512300993
acc_norm0.2878787878787879
acc_norm_stderr0.03225883512300993
hendrycksTest-high_school_government_and_politics
acc0.2538860103626943
acc_stderr0.0314102478056532
acc_norm0.2538860103626943
acc_norm_stderr0.0314102478056532
hendrycksTest-high_school_macroeconomics
acc0.2743589743589744
acc_stderr0.022622765767493207
acc_norm0.2743589743589744
acc_norm_stderr0.022622765767493207
hendrycksTest-high_school_mathematics
acc0.26296296296296295
acc_stderr0.026842057873833706
acc_norm0.26296296296296295
acc_norm_stderr0.026842057873833706
hendrycksTest-high_school_microeconomics
acc0.2647058823529412
acc_stderr0.028657491285071977
acc_norm0.2647058823529412
acc_norm_stderr0.028657491285071977
hendrycksTest-high_school_physics
acc0.304635761589404
acc_stderr0.03757949922943343
acc_norm0.304635761589404
acc_norm_stderr0.03757949922943343
hendrycksTest-high_school_psychology
acc0.28623853211009176
acc_stderr0.019379436628919968
acc_norm0.28623853211009176
acc_norm_stderr0.019379436628919968
hendrycksTest-high_school_statistics
acc0.25462962962962965
acc_stderr0.02971127586000535
acc_norm0.25462962962962965
acc_norm_stderr0.02971127586000535
hendrycksTest-high_school_us_history
acc0.23039215686274508
acc_stderr0.029554292605695053
acc_norm0.23039215686274508
acc_norm_stderr0.029554292605695053
hendrycksTest-high_school_world_history
acc0.2869198312236287
acc_stderr0.029443773022594693
acc_norm0.2869198312236287
acc_norm_stderr0.029443773022594693
hendrycksTest-human_aging
acc0.3811659192825112
acc_stderr0.03259625118416827
acc_norm0.3811659192825112
acc_norm_stderr0.03259625118416827
hendrycksTest-human_sexuality
acc0.1984732824427481
acc_stderr0.03498149385462472
acc_norm0.1984732824427481
acc_norm_stderr0.03498149385462472
hendrycksTest-international_law
acc0.3884297520661157
acc_stderr0.04449270350068382
acc_norm0.3884297520661157
acc_norm_stderr0.04449270350068382
hendrycksTest-jurisprudence
acc0.23148148148148148
acc_stderr0.04077494709252627
acc_norm0.23148148148148148
acc_norm_stderr0.04077494709252627
hendrycksTest-logical_fallacies
acc0.2331288343558282
acc_stderr0.03322015795776741
acc_norm0.2331288343558282
acc_norm_stderr0.03322015795776741
hendrycksTest-machine_learning
acc0.21428571428571427
acc_stderr0.03894641120044792
acc_norm0.21428571428571427
acc_norm_stderr0.03894641120044792
hendrycksTest-management
acc0.3300970873786408
acc_stderr0.04656147110012352
acc_norm0.3300970873786408
acc_norm_stderr0.04656147110012352
hendrycksTest-marketing
acc0.2905982905982906
acc_stderr0.029745048572674078
acc_norm0.2905982905982906
acc_norm_stderr0.029745048572674078
hendrycksTest-medical_genetics
acc0.29
acc_stderr0.04560480215720684
acc_norm0.29
acc_norm_stderr0.04560480215720684
hendrycksTest-miscellaneous
acc0.31545338441890164
acc_stderr0.016617501738763394
acc_norm0.31545338441890164
acc_norm_stderr0.016617501738763394
hendrycksTest-moral_disputes
acc0.2861271676300578
acc_stderr0.02433214677913413
acc_norm0.2861271676300578
acc_norm_stderr0.02433214677913413
hendrycksTest-moral_scenarios
acc0.2122905027932961
acc_stderr0.01367664468583173
acc_norm0.2122905027932961
acc_norm_stderr0.01367664468583173
hendrycksTest-nutrition
acc0.2875816993464052
acc_stderr0.02591780611714716
acc_norm0.2875816993464052
acc_norm_stderr0.02591780611714716
hendrycksTest-philosophy
acc0.2765273311897106
acc_stderr0.02540383297817961
acc_norm0.2765273311897106
acc_norm_stderr0.02540383297817961
hendrycksTest-prehistory
acc0.3117283950617284
acc_stderr0.025773111169630446
acc_norm0.3117283950617284
acc_norm_stderr0.025773111169630446
hendrycksTest-professional_accounting
acc0.26595744680851063
acc_stderr0.026358065698880592
acc_norm0.26595744680851063
acc_norm_stderr0.026358065698880592
hendrycksTest-professional_law
acc0.25684485006518903
acc_stderr0.011158455853098832
acc_norm0.25684485006518903
acc_norm_stderr0.011158455853098832
hendrycksTest-professional_medicine
acc0.1801470588235294
acc_stderr0.023345163616544855
acc_norm0.1801470588235294
acc_norm_stderr0.023345163616544855
hendrycksTest-professional_psychology
acc0.27941176470588236
acc_stderr0.018152871051538802
acc_norm0.27941176470588236
acc_norm_stderr0.018152871051538802
hendrycksTest-public_relations
acc0.3090909090909091
acc_stderr0.044262946482000985
acc_norm0.3090909090909091
acc_norm_stderr0.044262946482000985
hendrycksTest-security_studies
acc0.32653061224489793
acc_stderr0.030021056238440313
acc_norm0.32653061224489793
acc_norm_stderr0.030021056238440313
hendrycksTest-sociology
acc0.25870646766169153
acc_stderr0.030965903123573026
acc_norm0.25870646766169153
acc_norm_stderr0.030965903123573026
hendrycksTest-us_foreign_policy
acc0.32
acc_stderr0.04688261722621504
acc_norm0.32
acc_norm_stderr0.04688261722621504
hendrycksTest-virology
acc0.30120481927710846
acc_stderr0.0357160923005348
acc_norm0.30120481927710846
acc_norm_stderr0.0357160923005348
hendrycksTest-world_religions
acc0.32748538011695905
acc_stderr0.035993357714560276
acc_norm0.32748538011695905
acc_norm_stderr0.035993357714560276
truthfulqa_mc
mc10.2423500611995104
mc1_stderr0.01500067437357034
mc20.3859757929597962
mc2_stderr0.013898628036488968