2-3B Parameter LLMs — Page 49

gemma_unlearned_unbalance_gender_1e-7_1.0_0.5_0.75_epoch3

gemma_unlearned_unbalance_gender_1e-5_1.0_0.25_0.25_epoch1

gemma_unlearned_unbalance_gender_1e-7_1.0_1.0_1.0_epoch3

toy_goodharting_gemma-2-2b-it_fruits_vegetables_d_proxy_only_MC_20250403_092607_merged

merged-peft-final-model

gemma_unlearned_unbalance_gender_1e-6_1.0_0.05_0.15_epoch1

gemma_unlearned_unbalance_gender_1e-6_1.0_0.25_0.15_epoch1

gemma_unlearned_unbalance_gender_1e-7_1.0_0.5_0.25_epoch3

gemma_unlearned_unbalance_gender_1e-6_1.0_0.15_0.05_epoch2

gemma_unlearned_unbalance_gender_1e-7_1.0_0.15_0.05_epoch2

gemma_unlearned_unbalance_gender_1e-6_1.0_0.5_0.25_epoch1

gemma_unlearned_unbalance_gender_1e-7_1.0_0.25_0.5_epoch1

gemma-2-2b-dora

gemma_unlearned_unbalance_gender_1e-5_1.0_0.5_0.5_epoch3

gemma_unlearned_unbalance_gender_1e-6_1.0_1.0_1.0_epoch2

gemma_unlearned_unbalance_gender_1e-6_1.0_0.5_0.5_epoch2

Match-rigging_38

Match-rigging_35

gronger

DesiFamily-2B

Qwen2.5-Social-3B-NB-Chat

erpo-iclr-baseline-Qwen2.5-3B-dapo

PRM-qwen2.5-3b-alpacafarm-sft

qwen3_1.7b_sudoku_sft

manaba_gemma_2_2b

qwen3-custom-lm

qwen3_1.7b_easy_rl_final

meta-llama_Llama-3.2-3B-Instruct-GRPO-vanilla_G_4-checkpoint-88

meta-llama_Llama-3.2-3B-Instruct-GRPO-vanilla_G_4-checkpoint-292

qwen3_1.7b_easy_rl_fixed_gamma_1

qwen25-3b-qwq-aug-teacher-1e5

qwen3_1.7b_easy_rl_ours_adv_fixed_gamma_1_98_gem_ms_seq_is

qwen-3B-stego-4-codes

qwen-3B-stego-no-codes

qwen3_1.7b_easy_rl_ours_adv_fixed_geo_ms_seq_is

qwen3_1.7b_new_standard_B_sft_overfit_lr_5e_6__global_step_594

qwen3_1.7b_new_sudoku_one_action_new_sft_lr_5e_6

bartleby-llama-3.2-3b

self-debate-exp-Qwen2.5-3B-majority_fix_n4_l2048-DAPO_n8_bs256_long8-step200

qwen3_1.7b_sudoku_multi_act_new

agentic-futoshiki-qwen2.5-3B_SAS_SFT

qwen3_1.7b_rush_hour_one_move_final