albertge/llada-8b-dllm-registers-longarith-rl-r4

llada-8b-dllm-registers-longarith-rl-r4

Chunked GRPO on LongArithmetic starting from the mix60k R4 SFT main-triad checkpoint. Released at step 400, the step reported in tab:rl_results.

This checkpoint is part of the dLLM Registers project — register tokens as a bounded, trained, continuous channel for carrying decoding state across denoising windows in diffusion language models.

• —Paper section: tab:rl_results (LongArithmetic, registers arm)
• —Carry channel: continuous register tokens (channel_mode=registers, num_registers=4, tail_length=0)
• —Base model: GSAI-ML/LLaDA-8B-Base
• —Training data: LongArithmetic (5-8 operations, add/sub20, on-policy GRPO rollouts)
• —Training config: `diffu-grpo/amlt/mix60k_r4_longarith_expr5678_state_bonete54.yaml`
• —Chunk size: C = 128 tokens
• —Prompt dropout rate (Bernoulli per-trace CSG mask): 0.3
• —RL task: LongArithmetic
• —GRPO updates: 400
• —Initialized from: `albertge/llada-8b-dllm-registers-mix60k-r4`
• —Date uploaded: 2026-06-13

How to load

from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("albertge/llada-8b-dllm-registers-longarith-rl-r4", trust_remote_code=True)
tok = AutoTokenizer.from_pretrained("albertge/llada-8b-dllm-registers-longarith-rl-r4", trust_remote_code=True)

To use the carry channel correctly at inference, see the evaluator at `eval/eval.py` and the wrapper `eval/run_mix60k_full_eval.sh`. Key flags for this checkpoint: --num_registers 4 --channel_mode registers --tail_length 0 --usemasktokenforregisters

Repository

Training and eval code: https://github.com/lbertge/d1-registers

Citation

If you use this checkpoint, please cite the dLLM Registers paper:

@misc{dllm-registers-2026,
  title  = {Register Tokens for Unbounded Reasoning in Diffusion Language Models},
  author = {Albert Ge and collaborators},
  year   = {2026},
  note   = {Preprint},
  url    = {https://github.com/lbertge/d1-registers}
}