ThreadSumm: Summarization of Nested Discourse Threads Using Tree of Thoughts

TL;DR

ThreadSumm is a hierarchical LLM framework that improves nested discussion thread summarization by explicitly modeling discourse aspects, content units, and using a Tree of Thoughts search for coherent, multi-view summaries.

Contribution

It introduces a multi-stage, interpretable approach combining content planning, sentence ordering, and Tree of Thoughts search to enhance nested thread summarization.

Findings

Outperforms existing baselines in logical structure and aspect retention.

Achieves higher opinion coverage in nested discussion summaries.

Improves coherence and coverage through multi-proposal iterative refinement.

Abstract

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.