mh-ec2-ttomd-K2-delta

LoRA adapters for the K=2, delta cell of the coverage x debiasing table in the risk-sensitive preference-learning experiments (entropic risk, tau = 10).

Provenance

Trained locally by Max Horwitz.

Trained locally by Max on the UW Hyak cluster (SLURM job 40315482, finished 2026-09-20).

Full 4680-step run in one job, no resume.

What this run is

algorithm	online IPO (--alg oipo1), risk_egpo/tt_omd.py
coverage	K = 2 (--ypp_samples 2)
risk	entropic, tau = 10 (--risk entropic --risk_c 10.0)
debiasing	two-timescale bias correction, delta-method estimator
TT step size	gamma = 0.1
init	warm start from ipo-e-c10.0/checkpoint-936, 100 warmup steps
generation	64 max new tokens
seed	42
base model	vectorzhou/gemma-2-2b-it-alpaca-cleaned-SFT
dataset	PKU-Alignment/PKU-SafeRLHF
checkpoints	19 (250 ... 4680, every 250 steps)

Contents

checkpoint-<step>/ holds the LoRA adapter (adapter_model.safetensors, adapter_config.json) and tokenizer files. DeepSpeed resume state is not included.

from peft import PeftModel
from transformers import AutoModelForCausalLM
m = AutoModelForCausalLM.from_pretrained("vectorzhou/gemma-2-2b-it-alpaca-cleaned-SFT")
m = PeftModel.from_pretrained(m, "rat-lab/mh-ec2-ttomd-K2-delta", subfolder="checkpoint-4680")