| --- |
| license: apache-2.0 |
| language: |
| - en |
| base_model: dllm-hub/Qwen3-0.6B-diffusion-mdlm-v0.1 |
| tags: |
| - pytorch |
| - diffusion-language-models |
| - multi-agent |
| - latent-communication |
| - adapter |
| - arxiv:2610.28905 |
| --- |
| |
| # Hidden Messages: Tiny-A2D HM-Synth communication channel |
|
|
| Task-specific, supervised communication-channel weights accompanying |
| [Hidden Messages Between Diffusion Language Model Agents: Causal Evidence for Latent Communication](https://arxiv.org/abs/2610.28905), |
| accepted at the DiffuLM Workshop at NeurIPS 2026. |
|
|
| This is a communication adapter, **not a standalone language model**. The frozen |
| Tiny-A2D backbone is downloaded separately. The adapter consists of a sender-state |
| compressor and a gated residual receiver fuser, with four 128-dimensional message |
| slots. It was trained on HM-Synth distributed chain-lookup tasks. |
|
|
| ## Files and provenance |
|
|
| - `tiny_a2d_channel.pt`: original PyTorch state dictionary. |
| - `config.yaml`: original backbone revision, synthetic-task settings, channel |
| architecture, training settings, and evaluation settings. |
| - `LICENSE`: Apache-2.0 license for this release's channel weights and documentation. |
| - `THIRD_PARTY_LICENSES.md`: attribution for external dependencies and datasets. |
|
|
| These are the existing paper artifacts, mirrored without retraining or checkpoint |
| conversion from [the public code release](https://github.com/alanrbtx/hidden-messages-diffusion-agents/tree/1d1477b423ba4245b564840ce46aee039c79bcc8). |
| The checkpoint source is `checkpoints/tiny_a2d_channel.pt`. |
| No new inference or benchmark validation was performed for this Hub mirror. |
|
|
| ## Download |
|
|
| ```python |
| from huggingface_hub import hf_hub_download |
| |
| checkpoint_path = hf_hub_download( |
| repo_id="RobotLearningLab/hidden-messages-tiny-a2d", |
| filename="tiny_a2d_channel.pt", |
| ) |
| config_path = hf_hub_download( |
| repo_id="RobotLearningLab/hidden-messages-tiny-a2d", |
| filename="config.yaml", |
| ) |
| ``` |
|
|
| The download example does not run inference. The original `ReceiverChannel` class |
| is defined in |
| [`phase5_hm_synth_tiny.py`](https://github.com/alanrbtx/hidden-messages-diffusion-agents/blob/1d1477b423ba4245b564840ce46aee039c79bcc8/code/scripts/phase5_hm_synth_tiny.py), |
| with compressor and fusion modules in `code/src/hidden_messages/communication/`. |
| Use that architecture and the accompanying config to load the trusted state |
| dictionary with `torch.load(..., weights_only=True)`. The original driver includes |
| training and evaluation, not just a checkpoint-only inference CLI. This repository |
| does not implement Transformers `AutoModel.from_pretrained` or a hosted |
| text-generation endpoint. |
|
|
| Required backbone: `dllm-hub/Qwen3-0.6B-diffusion-mdlm-v0.1`, revision |
| `c8d24a3f4adaeef46881b450e1bf7d1005203bd7`. Use the pinned dependencies in the code |
| repository and a CUDA environment for the reported model workflow. The full |
| backbone weights are not included here. |
|
|
| ## Scope and limitations |
|
|
| Training uses union-context teacher states and paired matched-versus-deranged |
| supervision. Evaluation uses split private contexts. The adapter is task-specific; |
| it is not an emergent language, a privacy mechanism, or a demonstrated universal |
| cross-task protocol. It should not be assumed interchangeable with the Dream QASC |
| adapter. Independent agent trajectories, the designated receiver, and causal |
| message interventions are part of the evaluated setup. |
|
|
| For reported results and experimental scope, see the paper and |
| [`evidence/tiny_controlled_results.json`](https://github.com/alanrbtx/hidden-messages-diffusion-agents/blob/1d1477b423ba4245b564840ce46aee039c79bcc8/evidence/tiny_controlled_results.json). |
|
|
| ## Citation |
|
|
| ```bibtex |
| @misc{gazzaev2026hiddenmessages, |
| title={Hidden Messages Between Diffusion Language Model Agents: Causal Evidence for Latent Communication}, |
| author={Alan-Barsag Gazzaev and Alexey Gavrilov and Sergey Muravyov and Ilya Makarov}, |
| year={2026}, |
| eprint={2610.28905}, |
| archivePrefix={arXiv} |
| } |
| ``` |
|
|
