arxiv:2510.06557
Amirhossein Kazemnejad
kazemnejad
AI & ML interests
None yet
Recent Activity
upvoted a paper 2 days ago
Forecasting Downstream Performance of LLMs With Proxy Metrics upvoted a paper about 1 month ago
Structured Distillation of Web Agent Capabilities Enables Generalization upvoted a paper 7 months ago
Value Drifts: Tracing Value Alignment During LLM Post-TrainingOrganizations
models 11
kazemnejad/grpo_no_reasoning_sft4_complex_edits_v2_ckpt900_deepspeed
Updated
kazemnejad/imged_rl_grpo_no_reasoning_sft4_complex_edits_v2
Updated
kazemnejad/imged_rl_grpo_train_text_only_sft3_v2
Updated
kazemnejad/imged_rl_grpo_train_text_only_sft3_complex_edits
Updated
kazemnejad/imged_rl_grpo_no_reasoning_kl3e-4_lr3e-6
Updated
kazemnejad/reasoning_verbose-May1_got__256_1e-4_weightdecay0-stochastic-1.0-checkpoint-9600
Text Generation • 8B • Updated • 3
kazemnejad/imged_rl_grpo_full_no_reasoning_k8_v1
Updated
kazemnejad/Emu3-Base-SFT-got-Apr13_post_stochastic_256_lr1e-5-checkpoint-10800
Text Generation • 8B • Updated
kazemnejad/Janus-Pro-1B-unified-embed
2B • Updated
kazemnejad/emu3-grpo
Updated
datasets 0
None public yet