- Rating
- 1
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper proposes LD2LM that adapts drifting techniques to DreamReasoner8B. The method is directly built upon TokenDrift, while retaining the token-level cross-entropy loss and also combining with the multiscale feature drifting loss. Experiments are performed on top of DreamReasoner8B on code generation and math reasoning. Results show that drift-only training degrades performance while the CE loss and multiscale drifting loss preserves pass@1 and improves TPF.
Critical strengths
-
The paper adapts drifting loss to 8B-scale diffusion language model and evaluated on conditional generation tasks like coding and math reasoning.
-
The method is relatively simple and easy to implement.
Critical weaknesses
-
The experiment setting is troublesome and the results are not convincing. Table 1 caption explicitly lists "Math uses strict Math-Verify 0.8.0 and historical, non-matched runs" which is confusing. Most of the experiment results only show marginal improvement or even degradation, e.g., the method actually hurts LiveCodeBench performance compared to CE and even underperforms the starting checkpoint.
-
The novelty is limited and the method seems to be incremental to TokenDrift. The soft-token interface, FIFO queues, and the construction of the drifting field are already present in existing works. TokenDrift itself has also already explored a combination with CE loss. The experiment results also show that CE + feature KL often outperforms LD2LM and even matches LD2LM + tokenKL mostly.
Questions for authors
-
Why does CE continuation baseline has a score of 51.03 on LiveCodeBench in Table 1 but only 44.84 in Figure 3 and Table 2?
-
Why does LD2LM even hurt the performance on LiveCodeBench in Table 1 compared with the starting checkpoint?
-
In Table 2, CE + feature KL outperforms LD2LM on LiveCodeBench and ties on HE and MBPP. If this is already simple yet effective enough, why adding the FIFO queues and sophisticated designs in the drifting field?
-
The writings in the manuscript appear to be confusing and unnatural. A few examples below.
line 128: "provides the denoising foundation": what does denoising foundation mean?
line 138: "The distinction concerns the supervision space, rather than whether the model has access to sequence context": what does this mean?
line 207 and line 203 repeat the same sentence: "while the current remains differentiable"
line 300 unnatural narrative: "This is a history-length estimate, not a measured queue age. Queue capacity is a tested setting..."
The paper may require substantial rewriting.