{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Z2WFYDRM53RDQ5YMPSSEKZ643U","short_pith_number":"pith:Z2WFYDRM","schema_version":"1.0","canonical_sha256":"ceac5c0e2ceee238770c7ca44567dcdd22ff2b3aa1a97248ebc80c868e3a57a1","source":{"kind":"arxiv","id":"2501.04473","version":1},"attestation_state":"computed","paper":{"title":"When LLMs Struggle: Reference-less Translation Evaluation for Low-resource Languages","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Archchana Sindhujan, Constantin Orasan, Diptesh Kanojia, Shenbin Qian","submitted_at":"2025-01-08T12:54:05Z","abstract_excerpt":"This paper investigates the reference-less evaluation of machine translation for low-resource language pairs, known as quality estimation (QE). Segment-level QE is a challenging cross-lingual language understanding task that provides a quality score (0-100) to the translated output. We comprehensively evaluate large language models (LLMs) in zero/few-shot scenarios and perform instruction fine-tuning using a novel prompt based on annotation guidelines. Our results indicate that prompt-based approaches are outperformed by the encoder-based fine-tuned QE models. Our error analysis reveals tokeni"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2501.04473","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-08T12:54:05Z","cross_cats_sorted":[],"title_canon_sha256":"31a2db45cf7b6ffa19a645be95bc2591ea7929df7e0536129b0cdf5567ab4880","abstract_canon_sha256":"0b6cac9fcd0ced1c9b59d62278ea881aab3cc181c543326b031f1eb9d13b9cdb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:58:35.811328Z","signature_b64":"glK8bqUklLlJLD+wbLed5yXXOUWaHO+/MI0MlW01PXfvJeL4wIPZ5nVGe9rVovN88gwD+uAgJ+5QN41s/3eAAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ceac5c0e2ceee238770c7ca44567dcdd22ff2b3aa1a97248ebc80c868e3a57a1","last_reissued_at":"2026-07-05T09:58:35.810882Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:58:35.810882Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When LLMs Struggle: Reference-less Translation Evaluation for Low-resource Languages","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Archchana Sindhujan, Constantin Orasan, Diptesh Kanojia, Shenbin Qian","submitted_at":"2025-01-08T12:54:05Z","abstract_excerpt":"This paper investigates the reference-less evaluation of machine translation for low-resource language pairs, known as quality estimation (QE). Segment-level QE is a challenging cross-lingual language understanding task that provides a quality score (0-100) to the translated output. We comprehensively evaluate large language models (LLMs) in zero/few-shot scenarios and perform instruction fine-tuning using a novel prompt based on annotation guidelines. Our results indicate that prompt-based approaches are outperformed by the encoder-based fine-tuned QE models. Our error analysis reveals tokeni"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.04473","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2501.04473/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2501.04473","created_at":"2026-07-05T09:58:35.810937+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.04473v1","created_at":"2026-07-05T09:58:35.810937+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.04473","created_at":"2026-07-05T09:58:35.810937+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z2WFYDRM53RD","created_at":"2026-07-05T09:58:35.810937+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z2WFYDRM53RDQ5YM","created_at":"2026-07-05T09:58:35.810937+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z2WFYDRM","created_at":"2026-07-05T09:58:35.810937+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.12273","citing_title":"LLM-Based Evaluation of Low-Resource Machine Translation: A Reference-less Dialect Guided Approach with a Refined Sylheti-English Benchmark","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U","json":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U.json","graph_json":"https://pith.science/api/pith-number/Z2WFYDRM53RDQ5YMPSSEKZ643U/graph.json","events_json":"https://pith.science/api/pith-number/Z2WFYDRM53RDQ5YMPSSEKZ643U/events.json","paper":"https://pith.science/paper/Z2WFYDRM"},"agent_actions":{"view_html":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U","download_json":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U.json","view_paper":"https://pith.science/paper/Z2WFYDRM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.04473&json=true","fetch_graph":"https://pith.science/api/pith-number/Z2WFYDRM53RDQ5YMPSSEKZ643U/graph.json","fetch_events":"https://pith.science/api/pith-number/Z2WFYDRM53RDQ5YMPSSEKZ643U/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U/action/storage_attestation","attest_author":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U/action/author_attestation","sign_citation":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U/action/citation_signature","submit_replication":"https://pith.science/pith/Z2WFYDRM53RDQ5YMPSSEKZ643U/action/replication_record"}},"created_at":"2026-07-05T09:58:35.810937+00:00","updated_at":"2026-07-05T09:58:35.810937+00:00"}