AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

The Narrative Gap: Can LLMs help us Navigate Diverse Narratives Across Languages?

Nikhil Sharma, Kelly Marchisio, Kenton Murray, Ziang Xiao

NeurIPS 2026 · Large language models › Retrieval and knowledge · approach: Knowledge editing in LLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces a benchmark for testing whether LLMs can navigate diverse narratives about the same event across languages and regions.

Problem

No existing benchmark evaluates whether LLMs can navigate the diverse narratives that develop around the same event across languages and regions, which matters for information seeking about contested topics.

Approach

Argues that supporting diverse information seeking requires four core information navigation capabilities, and builds a benchmark to evaluate LLMs on narrative navigation across languages (details beyond the abstract are not stated).

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add an explicit diversity request (diverse prompt) to the query (ablation)14 LLMs on NarrativeBench · NarrativeBench narrative recallnarrative recall (%): 44.3 -> 49.2Improved
add an explicit diversity request (diverse prompt) (ablation)Gemini 2.5 Pro · NarrativeBench narrative recallnarrative recall (%): 70.6 -> 74.9Improved
ask for a specific narrative (inclusion) rather than excluding one14 LLMs · NarrativeBench NI vs NEnarrative F1 (%): NE 49.3 vs NI 19.9Hurt
use confirmatory instead of neutral query framing (ablation)14 LLMs · NarrativeBench BP/DP recallabsolute recall gap (pp): mean gap 3.23 pp; range 0.32 to 10.09Mixed
query in non-English language instead of English14 LLMs · NarrativeBench all tasksmean absolute gap vs English (pp): 8.74 pp query-language; 7.51 pp target narrative languageHurt
query in the same language as the source articles (ablation)source-language omission · NarrativeBench BP and DPsource-language omission rate (%): 50.4 -> 41.3Improved
enable medium reasoning (ablation)GPT-5.4-mini · NarrativeBench NInarrative F1 (%): 28.11 -> 39.37Improved
enable medium reasoning (ablation)GPT-5.4-mini · NarrativeBench NE, MNI, BP, DPF1/recall (pp): decrease of 7.38-15.94 ppHurt
minimum thinking settings and caps (ablation)Gemini Flash · NarrativeBench five tasksscore (pp): decrease of 8.09-10.59 ppHurt
higher translation quality (COMETKiwi) (ablation)benchmark task scores · NarrativeBench, 8 languagestask score: +0.17 pp per 0.01 COMETKiwi, CI [-0.15, 0.49]No clear effect
longer input context (per doubling) (ablation)task scores · NarrativeBench supplementary 1,720 casestask score (pp): +4.3 pp, 95% CI [-0.7, 9.3]No clear effect
share of post-knowledge-cutoff articles in context (ablation)nine models · NarrativeBenchrecall (pp): +4.77 pp, SE 4.84No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy