Downloads · 30 days
158
100% of all-time downloads
Snowball-67B-A2B-Math-RL-E6-Step20-Repaired
open-athena/Snowball-67B-A2B-Math-RL-E6-Step20-Repaired
Snowball-67B-A2B-Math-RL-E6-Step20-Repaired is a reinforcement learning model from open-athena. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. The card lists the license as other.
This is the exact checkpoint used for a reported row in the Snowball 67B-A2B math-RL experiment.
All-time downloads
158
Public
Parameters
67.1B
134 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors134 GB · 100%
Parameter types
How the weights are stored.
BF1667.1B · 100%
Model card
From the Hugging Face model README
Snowball-67B-A2B-Math-RL-E6-Step20-Repaired
This is the exact checkpoint used for a reported row in the Snowball 67B-A2B math-RL experiment.
Router state: SFT router-bias repaired; exact evaluated artifact. Closely named older laion/rl-snowball-* repositories may contain raw mutable-router exports that collapse under inference. Do not substitute them for this artifact. Models from this campaign are research checkpoints, not production releases; their utility is limited unless router-bias repair or frozen-router integrity is preserved.
- Arm:
E6 original - Step:
20 - Held-out AIME24 / MATH-500 / OlympiadBench:
26.00 / 76.80 / 22.67 - Selection: Headline E6 checkpoint in MATH_EVALS.md
- Source artifact:
s3://marin-us-east-02a/marin/diagnostics/snowball-bias-transplant-20260805/e6-s20-sftbias/policy/ - Experiment issue: https://github.com/marin-community/marin/issues/7786
- Evidence archive: https://huggingface.co/datasets/penfever/snowball-67b-a2b-math-rl-artifacts
The scores and evaluation caveats are recorded in MATH_EVALS.md in the evidence archive. Preserve config.json, the tokenizer files, and all shards named by model.safetensors.index.json together.
- Task
- Reinforcement Learning
- Type
- grug_moe
- License
- other
- Created
- Sep 20, 2026
- Updated
- Sep 20, 2026