ThreadSumm: Summarization of Nested Discourse Threads Using Tree of Thoughts
For researchers and practitioners dealing with complex online discussions, ThreadSumm addresses the problem of generating coherent, multi-viewpoint summaries from deeply nested threads, offering a structured approach that outperforms standard LLM summarizers.
ThreadSumm introduces a multi-stage LLM framework for summarizing deeply nested discussion threads by treating summarization as hierarchical reasoning over discourse aspects and Atomic Content Units, using a Tree of Thoughts search to optimize coherence and coverage. It achieves improved logical structure, aspect retention, and opinion coverage compared to baselines.
Summarizing deeply nested discussion threads requires handling interleaved replies, quotes, and overlapping topics, which standard LLM summarizers struggle to capture reliably. We introduce ThreadSumm, a multi-stage LLM framework that treats thread summarization as a hierarchical reasoning problem over explicit aspect and content unit representations. Our method first performs content planning via LLM-based extraction of discourse aspects and Atomic Content Units, then applies sentence ordering to construct thread-aware sequences that surface multiple viewpoints rather than a single linear strand. On top of these interpretable units, ThreadSumm employs a Tree of Thoughts search that generates and scores multiple paragraph candidates, jointly optimizing coherence and coverage within a unified search space. With this multi-proposal and iterative refinement design, we show improved performance in generating logically structured summaries compared to existing baselines, while achieving higher aspect retention and opinion coverage in nested discussions.