LLM reasoning / chain-of-thought
MoT (Merge of Thought)
Superseded baseline#468 of 772 most-superseded
Cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here
1 papers critique it · 0 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites MoT (Merge of Thought) as a baseline.
static parameter addition or averaging (e.g., Shen2025MergeofThoughtD) often induces conflicts within the parameter space in distillation tasks. This leads to destructive interference or ``cancellation effects'' among diverse supervision signals, resulting in suboptimal performance.
What to use instead
Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.
- Oct 15, 2025