GSM8K

GSM8K environment

GSM8K (grade-school math) environment with last-number verification, built on verifiers>=0.1.12. Single-turn: solve a word problem, ends-with-the-f...

GSM8K (grade-school math) environment with last-number verification, built on verifiers>=0.1.12. Single-turn: solve a word problem, ends-with-the-f...

GSM8K environment

GSM8K environment

GSM8K rubric-discovery environment: learn rubric_fn from (input, response, score) examples

GSM8K environment

Multi-turn math problem-solving environment where the model proposes Python code in a sandbox to compute and verify numerical answers.

Reward Hacking Sprint: does optimizing self-certainty (RLIF-style intrinsic reward) cause models to be confidently wrong on math? GSM8K, Llama-3.2-...

Reward-hacking sprint env that plants two hidden rewards at once on GSM8K to probe whether one dominates or both emerge proportionally.

Reward-hacking sprint env. Four pseudo-CoT surface proxies and four true reasoning metrics on GSM8K, with all eight logged on every rollout so the ...

Reward-hacking sprint env. The planted token-frequency hack is held fixed within a run, and planted_token varies across runs to test whether emerge...

Reward-hacking sprint env. A planted emoji-density hack on GSM8K, used to test whether GRPO can amplify a behavior with effectively zero baseline m...

Reward-hacking sprint env. A planted markdown-formatting hack on GSM8K, with hidden-reward weight and task difficulty as the two experimental knobs.

Reward-hacking sprint env. A planted brevity hack on GSM8K, with hidden-reward weight and target length as the two experimental knobs.

Reward-hacking sprint env. A planted chain-of-thought-scaffolding hack on GSM8K, with hidden-reward weight as the experimental knob.

Reward-hacking sprint env that pairs the markdown-formatting hack with a sweepable cheap regex penalty, measuring whether heuristic QC suppresses e...

Math reasoning - SingleTurnEnv + MathRubric

Just GSM8K with the added reward based on how shakespearean the model is.

GSM8K evaluation matching OLMES tulu_3_dev_no_safety methodology

AlphaZero-inspired MCTS environment for training LLMs through tree search and policy learning with teacher ensemble guidance and dual reward systems

Future House Aviary wrapper for verifiers - Scientific reasoning environments with tools

Future House Aviary wrapper for verifiers - Scientific reasoning environments with tools

Math group environment

GSM8K defend/concede environment for training calibrated self-assessment and sycophancy resistance

GSM8K with multi-reward support (correctness + length, optional gating)

GSM8K environment with a permissive last-number verifier

Encouraging reasoning model to produce more sensible thinking process by ask other model to understand and predict zeroshot answer from only reasoning text

An 860k-problem competition-math dataset with detailed solutions, the open community's go-to corpus for training math-specialized LLMs.

Allen AI's flagship open SFT mixture combining new persona-driven prompts with curated public data for post-training a frontier-quality instruct model.

Microsoft's "Evol-Instruct" recipe - automatically rewriting simple instructions into harder, more diverse ones using an LLM evolver.

Introduces GSM8K (8.5k grade-school math word problems) and shows that training a verifier to re-rank generated solutions outperforms simply fine-tuning on the dataset.

Introduces GSM8K (8.5k grade-school math word problems) and shows that training a verifier to re-rank generated solutions outperforms simply fine-tuning on the dataset.