List your company

Instruction Following Evaluation Suite

25.39Self-Instruct ScoreSFT
-4.17723.498911.17518.8511Jun 1, 2024Jul 14, 2024Aug 26, 2024Oct 9, 2024Nov 21, 2024Jan 3, 2025Feb 16, 2025

Evaluation results

MethodLinks
SFTModel=Phi3-mini-3.8B2025.0225.39——21.81——31.0929.1937.9741.07
SFTModel=Qwen2-7B2025.0222.69——21.42——30.3129.0737.3141.04
SFTModel=Llama-3-8B2025.0221.9——17——26.3627.0130.6635.23
FKLStudent=Qwen2-0.5B, Teacher=Qwen2-7B, Category=Same Tokenizer KD2025.0219.68——19.24——27.2927.4132.6737.46
SFTModel=Gemma-2-9B2025.0218.01——18.85——25.2326.7227.7434.83
CDMStudent=Qwen2-0.5B, Teacher=Phi3-mini-3.8B, Category=Cross Tokenizer KD2025.0216.55——16.38——25.125.4530.6636.47
MinEDStudent=Qwen2-0.5B, Teacher=Phi3-mini-3.8B, Category=Cross Tokenizer KD2025.0216.26——15.37——24.7225.5530.7635.69
MinEDStudent=Gemma-2-2B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0216.16——16.4——22.625.8325.9928.6
RKLStudent=Qwen2-0.5B, Teacher=Qwen2-7B, Category=Same Tokenizer KD2025.0216.15——16.62——25.3526.1530.3237.53
ULDStudent=Qwen2-0.5B, Teacher=Phi3-mini-3.8B, Category=Cross Tokenizer KD2025.0216.15——15.34——24.9326.4330.6336.07
SFTModel=OPT-6.7B2025.0215.71——15.82——24.0728.426.8733.56
FKLStudent=OPT-1.3B, Teacher=OPT-6.7B, Category=Same Tokenizer KD2025.0215.24——16.16——22.9225.3626.4731.38
SFTModel=Qwen2-0.5B2025.0215.17——15.22——24.0724.6630.3135
SFTModel=Gemma-2-2B2025.0214.94——16.89——22.4625.1225.2930.07
CDMStudent=Gemma-2-2B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0214.89——18.33——23.5526.1326.432
ULDStudent=Gemma-2-2B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0214.58——17.25——23.2326.1127.6930.53
CDMStudent=OPT-1.3B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0214.39——15.77——2326.1526.3332.33
FKLStudent=Gemma-2-2B, Teacher=Gemma-2-9B, Category=Same Tokenizer KD2025.0214.3——18.64——23.8226.5127.6132.06
SFTModel=OPT-1.3B2025.0214.26——14.81——22.4725.4825.8831.93
RKLStudent=Gemma-2-2B, Teacher=Gemma-2-9B, Category=Same Tokenizer KD2025.0213.8——18.64——22.2425.2623.729.79
ULDStudent=OPT-1.3B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0213.69——15.88——22.1825.4525.8230.07
RKLStudent=OPT-1.3B, Teacher=OPT-6.7B, Category=Same Tokenizer KD2025.0213.24——15.42——21.7725.0323.8631.27
MinEDStudent=OPT-1.3B, Teacher=Llama-3-8B, Category=Cross Tokenizer KD2025.0212.6——15.6——21.6925.2124.5130.52
Phased IFTBase Model=Llama-3 8B, Data=Alpaca-cleaned2024.065.753.91.150.624.795.433.97———
Phased IFTBase Model=Llama-2 7B, Data=Alpaca-cleaned2024.064.761.154.442.55.851.163.53———
Phased IFTBase Model=Llama-2 13B, Data=Alpaca-cleaned2024.064.562.985.8315.626.9110.856.49———
Phased IFTBase Model=Llama-2 7B, Data=Alpaca2024.064.176.425.566.889.3114.347.26———
Phased IFTBase Model=Llama-3 8B, Data=Alpaca2024.063.059.679.527.56.3212.67.64———
Phased IFTBase Model=Llama-2 13B, Data=Alpaca2024.061.982.757.7811.8811.4416.287.35———
Phased IFTBase Model=Llama-3 70B, Data=Alpaca2024.061.6510.15.235.842.347.875.23———
Phased IFTBase Model=Llama-2 70B, Data=Alpaca-cleaned2024.060.46.428.3312.510.914.737.59———
Phased IFTBase Model=Mistral 7B, Data=Alpaca2024.06-0.65.966.1116.257.9812.026.3———
Phased IFTBase Model=Llama-2 70B, Data=Alpaca2024.06-3.042.577.2713.464.102.82———