sqa-grpo-temp12-step700

GRPO + temperature 1.2 baseline on ScienceQA, trained with GRPO on Qwen/Qwen2.5-Math-1.5B.

Selected as best validation pass@6 for this arm (rank 2).

Exploration via the sampling distribution. Trained at rollout temperature 1.2 but VALIDATED at 1.0 like every other arm, so its curve stays comparable.

Do not apply a chat template

Trained on raw prompt text. verl's RLHFDataset has apply_chat_template=False and it was never enabled. Applying Qwen2.5-Math's chat template at inference creates a train/eval mismatch measured at roughly 19 points of pass@1 on a sibling task.

from vllm import LLM, SamplingParams
llm = LLM(model="sandeep123/sqa-grpo-temp12-step700", dtype="bfloat16", max_model_len=1536)
params = SamplingParams(n=6, temperature=1.0, top_p=1.0, top_k=-1, max_tokens=1024)
out = llm.generate([prompt_text], sampling_params=params)   # raw string, not llm.chat()

Validation metrics at this checkpoint

metric	value
pass@1	0.8145
pass@6	0.9648
step	700

Answer extraction (pre-registered). An answer is the content of the final \boxed{}; if absent, the last standalone A-E token. Responses with no extractable answer are scored incorrect, and all K rollouts stay in the denominator. This is ScienceQA's answer-choice accuracy, reported as "sampled answer accuracy (pass@1)".

Validation uses 256 held-out prompts, K=6, temperature 1.0, seed 42 -- pinned in code so every arm, including the temperature-1.2 arm, is scored under identical decoding.

Settings (identical across all baseline arms)

dataset	ScienceQA (scienceqa_boxfix)
epochs / steps	25 / 1250
batch / rollouts	128 prompts, K=6
learning rate	1e-6 constant
KL (in-reward)	0.01
max prompt / response	512 / 1024 tokens
format reward	0.03, constant, no decay
seed	42

Arm-specific: entropy_coeff=0.0, clip=0.2/0.2, rollout temperature=1.2.

Note on checkpoint selection

Quality-optimal and diversity-optimal checkpoints differ substantially: for these runs the best-pass@1 checkpoint lands near step 1000-1200 while the best-pass@6 checkpoint is near step 200-500. Both are published so that diversity results are not reported from a checkpoint chosen purely for accuracy.