8B, 2B, 4B) using an autoregressive masked language modeling objective.
Multimodal decision models output structured choices in a single pass
Intern-Decision uses autoregressive masked language modeling to produce candidate probabilities and decisions from text and images, achieving high accuracy and low latency.
Chinese Tech
Tong Zhu · Zhe Xu · Zefeng He · Yunkai Chai · Xuyang Hu · Dehong Zheng · +4 more
Shanghai AI Lab · Shanghai Jiao Tong University · Fudan University · Nanyang Technological University
Research Digest··3 min read
Thread:RL for Tool Agents
8B, 2B, and 4B parameters) that reformulate structured decision making as next-token prediction over masked fields in a predefined JSON skeleton, enabling multi-field scoring in a single causal forward pass.
Why this paper
From Shanghai AI Lab and 3 others · Part of RL for Tool Agents, now 26 papers
In one line
Intern-Decision makes structured multimodal decisions in a single forward pass, achieving high accuracy and low latency.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ✓Compute or model size stated (hardware RTX 4090)
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks (4 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§