Pith. sign in

REVIEW 8 cited by

BooookScore: A systematic exploration of book-length summarization in the era of LLMs

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.00785 v4 pith:QH7HFXFE submitted 2023-10-01 cs.CL cs.AIcs.LG

BooookScore: A systematic exploration of book-length summarization in the era of LLMs

classification cs.CL cs.AIcs.LG
keywords booookscorellmsbook-lengthevaluationhumansummariesmodelsannotations
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Summarizing book-length documents (>100K tokens) that exceed the context window size of large language models (LLMs) requires first breaking the input document into smaller chunks and then prompting an LLM to merge, update, and compress chunk-level summaries. Despite the complexity and importance of this task, it has yet to be meaningfully studied due to the challenges of evaluation: existing book-length summarization datasets (e.g., BookSum) are in the pretraining data of most public LLMs, and existing evaluation methods struggle to capture errors made by modern LLM summarizers. In this paper, we present the first study of the coherence of LLM-based book-length summarizers implemented via two prompting workflows: (1) hierarchically merging chunk-level summaries, and (2) incrementally updating a running summary. We obtain 1193 fine-grained human annotations on GPT-4 generated summaries of 100 recently-published books and identify eight common types of coherence errors made by LLMs. Because human evaluation is expensive and time-consuming, we develop an automatic metric, BooookScore, that measures the proportion of sentences in a summary that do not contain any of the identified error types. BooookScore has high agreement with human annotations and allows us to systematically evaluate the impact of many other critical parameters (e.g., chunk size, base LLM) while saving $15K USD and 500 hours in human evaluation costs. We find that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher BooookScore than those generated by open-source models. While LLaMA 2 falls behind other models, Mixtral achieves performance on par with GPT-3.5-Turbo. Incremental updating yields lower BooookScore but higher level of detail than hierarchical merging, a trade-off sometimes preferred by annotators.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing

    cs.LG 2026-01 unverdicted novelty 7.0

    HER trains LLMs on reverse-engineered reasoning data and human preference rewards to improve cognitive persona simulation, reporting 30-point gains on CoSER and 15% on Minimax benchmarks over Qwen3-32B.

  2. StoryLens: Preference-Aligned Story Rewriting via Context-Aware Narrative Enrichment

    cs.CL 2026-05 unverdicted novelty 6.0

    StoryLens creates a benchmark and models for context-enriched story rewriting that better matches reader preferences than style transfer alone.

  3. WiCER: Wiki-memory Compile, Evaluate, Refine Iterative Knowledge Compilation for LLM Wiki Systems

    cs.CL 2026-05 conditional novelty 6.0

    WiCER iteratively diagnoses and repairs fact loss during wiki compilation for LLMs, recovering 80% of quality lost in blind distillation across 17 domains while cutting catastrophic failures by 55%.

  4. Whose Story Gets Told? Positionality and Bias in LLM Summaries of Life Narratives

    cs.CL 2026-04 unverdicted novelty 6.0

    A proposed pipeline shows LLMs introduce detectable race and gender biases when summarizing life narratives, creating potential for representational harm in research.

  5. ThreadSumm: Summarization of Nested Discourse Threads Using Tree of Thoughts

    cs.CL 2026-04 unverdicted novelty 6.0

    ThreadSumm improves structured summarization of nested discourse threads by combining LLM-based aspect and content unit extraction with sentence ordering and Tree of Thoughts search for better coherence and opinion coverage.

  6. CheckSupport: A Local LLM-Powered Tool for Automated Manuscript Submission Checklist Selection and Completion

    cs.DL 2026-05 conditional novelty 5.0

    CheckSupport uses local LLMs with staged prompting to recommend and complete reporting checklists for manuscripts, reporting 90% recommendation accuracy and 88% item-level completion accuracy on peer-reviewed papers.

  7. ResearchPulse: Building Method-Experiment Chains through Multi-Document Scientific Inference

    cs.CL 2025-09 conditional novelty 5.0

    ResearchPulse extracts motivation-method chains and experimental trends from related papers, rendering them as mind maps and line charts, and releases a 100-cluster benchmark; the reported '7B beats GPT-4o' result is ...

  8. A Tree-of-Thoughts Inspired Hybrid Approach for Legal Case Judgement Summarization using LLMs

    cs.CL 2026-06 unverdicted novelty 3.0

    A tree-of-thoughts inspired hybrid extractive-abstractive LLM prompt yields better legal case judgment summaries than standard extractive or abstractive prompts.