{"total":11,"items":[{"citing_arxiv_id":"2607.00890","ref_index":101,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"MultiSynt/MT: Trillion-Token Multi-Parallel Pre-Training Data Translated Across 36 Languages","primary_cat":"cs.CL","submitted_at":"2026-07-01T12:55:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MultiSynt/MT supplies 4.8 trillion translated tokens in 36 languages from 100B English tokens, letting LLMs match native-data baselines with 72% fewer tokens and beat them by 15% at equal budget.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.26466","ref_index":74,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Soft Token Alignment for Cross-Lingual Reasoning","primary_cat":"cs.CL","submitted_at":"2026-06-25T00:01:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SOLAR aligns soft-token probability mixtures across languages in embedding space during SFT and raises multilingual reasoning accuracy by up to 17.7 points over the base model.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.25821","ref_index":31,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"SARA: Unlocking Multilingual Knowledge in Mixture-of-Experts via Semantically Anchored Routing Alignment","primary_cat":"cs.CL","submitted_at":"2026-06-24T13:36:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SARA aligns internal routing distributions in MoE layers to high-resource semantic anchors via symmetric JS divergence, improving low-resource language performance by 0.8-1.2% over standard instruction tuning on Global-MMLU.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.07167","ref_index":11,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"UrduMMLU: A Massive Multitask Benchmark for Urdu Language Understanding","primary_cat":"cs.CL","submitted_at":"2026-06-05T11:35:27+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"UrduMMLU is a new native-source MCQ benchmark for Urdu that reveals top LLMs reach only ~90% accuracy with large gaps on region-specific humanities content.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.04694","ref_index":13,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"DuDi: Dual-Signal Distillation with Cross-Lingual Verbalizer","primary_cat":"cs.CL","submitted_at":"2026-06-03T10:23:05+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"DuDi is a dual-signal distillation method with cross-lingual verbalizer that improves multilingual SLM performance on SEA languages and outperforms baselines on SEA-HELM.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02465","ref_index":19,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Learning When to Translate for Multilingual Reasoning","primary_cat":"cs.CL","submitted_at":"2026-06-01T16:37:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Luar is a reinforcement learning method enabling reasoning language models to decide when to invoke English translation for improved multilingual reasoning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.01800","ref_index":53,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Multilinguality of Large Language Models From a Structural Perspective","primary_cat":"cs.CL","submitted_at":"2026-06-01T07:18:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Low-resource languages are structurally more different from English in LLMs than high- or mid-resource ones, and language-specific post-training alters structures while preserving inter-language relationships.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.28163","ref_index":42,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"DEPART: DEcomposing PARiTy across Multilingual LLMs","primary_cat":"cs.CL","submitted_at":"2026-05-27T08:45:48+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A Bayesian framework decomposes mLLM variance, showing language features explain 79-92% of language identity variance and that model identity vs. benchmark-model interactions dominate differently for understanding versus reasoning tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.22137","ref_index":17,"ref_count":2,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Cross-Lingual Consensus: Aligning Multilingual Cultural Knowledge via Multilingual Self-Consistency","primary_cat":"cs.CL","submitted_at":"2026-05-21T08:11:01+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A multilingual self-consistency plus self-critique method raises cultural alignment scores on English queries by 5.03% on the BLEnD benchmark using only self-generated data.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.13412","ref_index":52,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"LLMs as annotators of credibility assessment in Danish asylum decisions: evaluating classification performance and errors beyond aggregated metrics","primary_cat":"cs.CL","submitted_at":"2026-05-13T12:07:47+00:00","verdict":"ACCEPT","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"LLMs can provide cost-effective annotation of credibility in Danish asylum texts but produce inconsistent errors that vary by model and prompt, requiring checks beyond single-model accuracy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.11632","ref_index":49,"ref_count":1,"confidence":0.88,"is_internal_anchor":false,"paper_title":"Macro: Enhancing Multilingual Counterfactual Explanations through Alignment-as-Preference Optimization","primary_cat":"cs.CL","submitted_at":"2026-05-12T06:56:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Macro uses DPO on composite preference pairs to raise validity of multilingual self-generated counterfactual explanations by 12.55% on average over chain-of-thought while preserving minimality.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}