Lost in Interpretation: The Plausibility-Faithfulness Trade-off in Cross-Lingual Explanations

• URL: http://arxiv.org/abs/2605.19274v1
• Date: Tue, 19 May 2026 02:44:48 GMT
• Title: Lost in Interpretation: The Plausibility-Faithfulness Trade-off in Cross-Lingual Explanations
• Abstract summary: We find that English explanations frequently produce fluent but loosely anchored rationales.<n>For socially nuanced classification, English pivots also fail to preserve pragmatic cues.
• Score: 7.934534873467547
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: LLMs deployed multilingually are often audited via English explanations for non-English inputs. We evaluate extractive explanations ''where the model identifies input token spans as evidence alongside a generated rationale'' and uncover a systematic trade-off: English-pivot explanations can achieve higher span agreement with human rationales while their evidence becomes less causally grounded in the model's prediction, as measured by both comprehensiveness and sufficiency. Across 3 tasks, 5~languages, and 2~multilingual LLM families, we find that English explanations frequently produce fluent but loosely anchored rationales, with comprehensiveness degrading by up to 5.7x relative to native-language conditions - even as task accuracy remains stable across settings. For socially nuanced classification, English pivots also fail to preserve pragmatic cues, reducing both faithfulness and span agreement. We recommend auditing explanations in the input language, reporting multi-faceted faithfulness metrics beyond lexical overlap, and treating English rationales as communication summaries rather than faithful decision traces.

Related papers

• The Interlingua Hypothesis: LLMs Translate via a Latent Task-agnostic Feature Space [19.151700851322897]
  Large language models (LLMs) have recently demonstrated improved machine translation performance over strong supervised baselines.<n>Motivated by recent interpretability findings, we propose the interlingua hypothesis.<n>The hypothesis holds that language models translate by reading a source sentence into a latent feature space, and generate a target sentence by reading from the latent feature space.
  arXiv  Detail & Related papers  (2026-09-01T00:30:22Z)
• LANG: Reinforcement Learning for Multilingual Reasoning with Language-Adaptive Hint Guidance [77.58408743830314]
  Reinforcement learning has proven effective for enhancing multi-step reasoning in large language models.<n>But its benefits have not fully translated to multilingual contexts.<n>We develop a novel framework that leverages language-conditioned hints to guide exploration in non-English reasoning tasks.
  arXiv  Detail & Related papers  (2026-05-21T14:47:52Z)
• Plausibility as Commonsense Reasoning: Humans Succeed, Large Language Models Do not [0.0]
  We test whether large language models integrate world knowledge with syntactic structure in a human-like, structure-sensitive way during ambiguity resolution.<n>We construct ambiguous items that keep the syntactic configuration fixed and ensure both parses remain pragmatically possible.<n>In a speeded forced-choice comprehension experiment, humans show a large, correctly directed plausibility effect.<n>Across models, plausibility-driven shifts are weak, unstable, or reversed.
  arXiv  Detail & Related papers  (2026-04-06T16:27:06Z)
• When Meanings Meet: Investigating the Emergence and Quality of Shared Concept Spaces during Multilingual Language Model Training [57.230355403478995]
  We investigate the development of language-agnostic concept spaces during pretraining of EuroLLM.<n>We find that shared concept spaces emerge early and continue to refine, but that alignment with them is language-dependent.<n>In contrast to prior work, our fine-grained manual analysis reveals that some apparent gains in translation quality reflect shifts in behavior.
  arXiv  Detail & Related papers  (2026-01-30T11:23:01Z)
• Align to the Pivot: Dual Alignment with Self-Feedback for Multilingual Math Reasoning [71.4175109189942]
  We present Pivot-Aligned Self-Feedback Multilingual Reasoning (PASMR)<n>This approach designates the model's primary language as the pivot language.<n>It establishes a cross-lingual self-feedback mechanism without relying on external correct answers or reward models.
  arXiv  Detail & Related papers  (2026-01-25T03:20:00Z)
• Large Reasoning Models Are (Not Yet) Multilingual Latent Reasoners [48.68444770923683]
  Large reasoning models (LRMs) achieve strong performance on mathematical reasoning tasks.<n>LRMs often arrive at the correct answer before completing these textual reasoning steps.<n>This phenomenon has been explored in English, but its multilingual behavior remains largely unknown.
  arXiv  Detail & Related papers  (2026-01-06T13:20:17Z)
• Beg to Differ: Understanding Reasoning-Answer Misalignment Across Languages [43.36259715282423]
  We analyze 65k reasoning traces from GlobalMMLU questions across 6 languages and 6 frontier models.<n> Reasoning traces in non-Latin scripts show at least twice as much misalignment between their reasoning and conclusions than those in Latin scripts.
  arXiv  Detail & Related papers  (2025-12-27T21:55:21Z)
• Can LLMs Faithfully Explain Themselves in Low-Resource Languages? A Case Study on Emotion Detection in Persian [0.0]
  Large language models (LLMs) are increasingly used to generate self-explanations alongside their predictions.<n>This study evaluates the faithfulness of LLM-generated explanations in the context of emotion classification in Persian.
  arXiv  Detail & Related papers  (2025-11-24T21:29:15Z)
• Think Natively: Unlocking Multilingual Reasoning with Consistency-Enhanced Reinforcement Learning [85.7304930030649]
  We propose M-Thinker, which is trained by a Language Consistency reward and a Cross-lingual Thinking Alignment reward.<n>M-Thinker achieves nearly 100% language consistency and superior performance on two multilingual benchmarks.
  arXiv  Detail & Related papers  (2025-10-08T17:55:02Z)
• When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners [111.50503126693444]
  We show that language-specific ablation consistently boosts multilingual reasoning performance.<n>Compared to post-training, our training-free ablation achieves comparable or superior results with minimal computational overhead.
  arXiv  Detail & Related papers  (2025-05-21T08:35:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.