After working on cleaning the text for modelling, I've found some transliteration/translation pairs that seem weird and was hoping to get some clarification on if this stuff will be seen in the hidden test set or if it has been cleaned in a certain way that we should replicate.
\nI'm bolding the text I'm questioning in the following examples. These examples were not cleaned.
\nExample
\nTransliteration: KIŠIB šu-(d)EN.LÍL DUMU šu-ku-bi-im KIŠIB ṣí-lu-lu DUMU ú-ku i-nu-mì i-dí-a-bu-um a-wa-sú iq-bi-ú 10 ma-na KÙ.BABBAR a-na ša-lim-a-šùr i-dí-in um-ma šu-ut-ma i-ṣí-ba-at KÙ-pì-a li-il₅-qé
\nTranslation: Seal of Šu-Illil son of Šu-Kūbum, seal of Ṣilūlu son of Uku. When Iddin-abum spoke his will, he gave 10 minas ofדsilver to Šalim-Aššur. He said: He may take it from the interest on my silver.\"\"
\nExample
\nTransliteration: um-ma šu-ku-tum-ma a-na IŠTAR-lá-ma-sí ù ni-ta-aḫ-šu-šar qí-bi₄-ma mì-šu ša ta-áš-pu-ra-ni-ni um-ma a-tí-na-ma É-tum a-na lá be-tim i-tù-ar a-pu-tum a-na en-um-a-šùr i-xx-ni-ma e ší-na ga x ša lá ta-ḫa-dì-ri a-na IŠTAR-lá-ma-sí qí-bi₄-ma šu-ma a-ḫa-tí a-ta li-ba-am dì-ni-ší-im lá ta-ḫa-da-ar a-na ni-ta-aḫ-šu-šar qí-bi₄-ma TÚG-pì-ri-kà-ni ša e-zi-bu na-pí-ší-šu-nu ù ṭup-pu-ú lu ša-ṣú-ru pì-ri-kà-nu ša ma-tí ù tí-bu-lá ma-a-x iš-ta-ú-mu-ni a-dí en-um-a-šùr i-lá-kà-ni a ma-ma-an lá tù-šé-ri x GÍN KÙ.BABBAR (d)UTU-tap-pá-i ub-lá-ki-im 1 GÍN KÙ.GI ù x GÍN KÙ.BABBAR i-ku-pì-a ub-lá-ki-im
\nTranslation: From Šukkutum to Ištar-lamassī and Nitahšušar: Why is that you (fem. plur.) have written me, saying: The house is no longer a house.\" Urgent, to Ennam-Aššur … Do not fear!. To Ištar-lamassī: If you are truly my sister, then encourage her. Do not fear. To Nitahšušar: Air the -textiles that I left. Also, the tablets should be guarded. The which Mati? and Tibula … have bought do not release (them) to anyone before Ennam-Aššur arrives. Šamaš-tappā'ī brought you x shekels of silver. Ikūn-pīya brought you 1 shekel of gold and x shekels of silver. \"
\nExample
\nTransliteration: 30 ma-na KÙ.BABBAR ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a-na um-mì-a-ni-a ša ki-ma i-a-tí ù ḫi-na-a šu-ku-bu-um ú-bi₄-il₅ IGI a-mur-(d)UTU DUMU mì-šar-ra-bí IGI a-mur-(d)UTU DUMU ḫa-ri-im-tim 2 ma-na KÙ.BABBAR i-a-am 1 ma-na KÙ.BABBAR ša ḫi-na-a a pá-ni ILLAT-tim 1.5 ma-na KÙ.BABBAR ša um-mì-iš-ḫa-ra 15 GÍN KÙ.KI ša lá-ma-sí-tim ú a-šur-iš-tí-kál ŠÀ.BA 0.5 ma-na 5 GÍN ša lá-ma-sí-tim 0.33333 ma-na 5 GÍN KÙ.BABBAR ša ik-ri-bi-a mì-ma a-nim ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a ša ki-ma i-a-tí šu-ku-bu-um ú-bi₄-il₅ IGI lá-qé-ep DUMU i-zi-za-am-ì-lí IGI a-šur-na-da
\nTranslation: 30 minas of silver, its excise [added], his transport fee paid, under my seals, Šu-Kūbum brought to my investors and my representatives and to hinnāya. (Entrusted to him) in the presence of Amur-Šamaš, son of Mīšar-rabi, of Amur-Šamaš, son of harimtum. 2 minas of silver belonging to me; 1 mina of silver belonging to hinnāya, (sent) to meet the caravan; 1.5 mina of silver of Ummī-Išhara; 15 shekels of gold of Lamassutum and Aššuriš-tikal, thereof 35 shekels of silver of Lamassutum (and?) 25 shekels of [silver] being [my?] temple funds - all this, its import tax [added], his transport fee [paid], under my seals, Šu-Kūbum brought to my representatives. In the presence of Lā-qēp, son of Izzizam-ilī, of Aššur-nādā.
\nExample (it's the fractions)
\nTransliteration: 2 né-pí-šu 15 ma-na.TA ú iš-tí-in né-pí-šu-um 10 ma-na ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu-ú ŠU.NÍGIN 42.33333 ma-na KÙ.BABBAR ṣa-ru-pá-am ku-nu-ki-a a-na a-lu-wa ù e-ni-ša-ri-im áp-qí-id-ma a-na a-lim(ki) a ma-lá tí-ir-tí-šu a-ṣé-er ša-lim-a-šùr ú-šé-bi-il₅-šu-nu a-ḫa-ma 13.33333 ma-na URUDU SIG₅ a-na ga-am-ri-šu-nu ù 5 GÍN KÙ.BABBAR a-na ú-ku-ul-tí-šu-nu a-dí-in IGI ili₅-ba-ni DUMU ba-ší-lam IGI a-ḫu-qar DUMU zu-ur-zu-ur IGI tù-ra-am-ì-lí DUMU e-dí-na-a-šùr a-ḫa-ma 10.33333 ma-na 3.5 GÍN KÙ.BABBAR ṣa-ru-pá-am x tum ni x x x 0.5 GÍN ṣí-ba-tim ša i-na ṣé-ri-a il₅-qé-ú-ni a-na ḫu-bu-li-šu a-na kà-ri-im wa-aḫ-šu-ša-na áš-qúl ṣí-ba-at KÙ.GI ša ší-ip-kà-at a-šur-bé-el-a-wa-tim i-ší-tù
\nTranslation: 2 packages of 15 minas each plus a single package of 10 minas, its import duty added, its transport tariff paid - in all 42 ⅓ minas of refined silver under my seal, I entrusted to Aluwa and Enišārum, and I sent them to the City to Šalim-Aššur in accordance with his orders. Furthermore, I paid 13 ⅓ minas of good copper for their expenses and 5 shekels of silver for their food. Witnessed by Ilī-bāni son of Baši-ilum, by Ahu-qar son of Zurzur, by Tūram-ilī son of Eddin-Aššur. Furthermore, 10 ⅓ minas 3 ½ shekels of refined silver …, … ½, shekel interest that they took on my account, I paid for his debt to the Wahšušana colony. The interest on the gold that remained of Aššur-bēl-awātim's investment.
\nExample
\nTransliteration: 4 GÍN a-na ší-iṭ-ri-im ša pu-ki-im 1.3333300000000001 GÍN a-na e-re-qí-im qá-nu-e áš-qúl 1.66666 GÍN a sú-ba-ri-im áš-qúl 0.33333 ma-na KÙ.BABBAR a-na a-bar-ni-im áš-qúl 0.16666 GÍN a-na um-ṣí-im 0.25 GÍN a šu-um-ki xxx x na-ru-uq GIG xxxx […] xx GÍN a-na … ḫa-áš-lá-tim xxx a-wa-ar-nu-a-lim 0.66666 GÍN a-na e-ṣé áš-qúl 0.66666 GÍN a pá-e ú-ša-qí-il₅ 1 bi-il₅-té-en 0.5 GÍN áš-qúl 0.25 GÍN a-na na-pá-ḫi-im 95 ki-ra-tim a-na x 0.25 GÍN.TA ù 7.5 ŠE.TA 0.5 GÍN a-na xxxx 0.16666 GÍN a-na xxxx a-na xxxx
\nTranslation: I paid 4 shekels for a scarf of -weave, 1.3333 shekel for a wagonload of reed; I paid 1.6666 shekel for ; I paid 0.3333 mina of silver for an Abarnian textile, 1 / 6 shekel for a piece of dried meat, 1 / 4 shekel for onions, [x shekels for] [x] bags of wheat… …;[I paid … … … … … …] shek[els for] ?, [… … … …] for a … … … …; I paid 0.6666 shekel for firewood; 0.6666 shekels I paid for chaff, for a double load I paid 0.5 shekel; 1 / 4 shekel for the blacksmith. I supplied 95 drinks to .. at 52.5 grains (of silver) each, 0.5 shekel for … … … 1 / 6 shekel for … for … … … … … … …
\nExample
\nTransliteration: 1.25 GÍN KÙ.BABBAR i-na li-bi₄ lu-lu
\nTranslation: 1 1 / 4 shekel of silver is owed by Lulu.
\nExample
\nTransliteration: […] x šu-a-x-… um-ma … … … … en-nam-a-šur … bi-ma … tám-kà-ri-im … ḫu-a … … KÙ.BABBAR IGI e-ni-iš-ta-ru-um … ni AN.NA-ki … ší-a-ni … … ma x ku-un-kà-ma a-na e-ni-iš-ta-ri-im dí-na-ma iš-tí ba-tí-qí-im ša a-na pá-ni ILLAT-tim i-lá-kà-ni lá i-sà-ḫu-ur ṭù-ur-da-ni-šu a-na-kam we-da-ku a-ḫu-a a-tù-nu e-ni-iš-ta-ra-am lá i-sà-ḫu-ur ṭù-ur-da-ni-šu mì-li-ik lu-qú-tim x x bi-ri-im mì-il₅-kà sà-ḫe-er-tám ša 5 GÍN KÙ.BABBAR x x nim a-na ší-im … … … … mì-li-ik-šu-nu mì-il₅-kà
\nTranslation: [Obverse too broken for translation]
\nExample
\nTransliteration: lu-lu en-nam-a-šur ù a-bu-um-ša-lim ša ki-ma a-šùr-DU₁₀ DUMU e-lá-ma a-na a-šur-DU₁₀ DUMU a-lá-ḫi-im iṣ-bu-tù-ni-a-tí-ma um-ma šu-nu-ma ší-im 21 GÚ 9 ma-na URUDU ma-sí-im ša a-šur-DU₁₀ DUMU e-lá-ma ša i-na É-tí-kà i-ni-id-ú a-na ší-mì-im ta-dí-šu um-ma a-šur-DU₁₀-ma a-dí-šu-ma 17.83333 ma-na 2 GÍN KÙ.BABBAR ší-im-šu na-áš-a-ku eq-lam a šu-mì a-šur-DU₁₀ DUMU e-lá-ma KÙ.BABBAR e-tí-iq a-na a-wa-tim a-ni-a-tim kà-ru-um bu-ru-uš-ḫa-tum i-dí-ni-a-tí-ma ší-bu-tí-ni IGI GÍR ša a-šur ni-dí-in IGI i-ku-pì-a IGI ú-ṣú-ur-ša-a-šur
\nTranslation: Lullu,
\nThere's a lot more, like transliterations with lots of gaps but no '…' in the translation or the opposite scenario. I may share a notebook that contains more examples if there's any interest. Any help would be appreciated.
", "messages": [ { "id": "3382583", "postDate": "12/27/2025 22:45:44", "content": "After working on cleaning the text for modelling, I've found some transliteration/translation pairs that seem weird and was hoping to get some clarification on if this stuff will be seen in the hidden test set or if it has been cleaned in a certain way that we should replicate.
\nI'm bolding the text I'm questioning in the following examples. These examples were not cleaned.
\nExample
\nTransliteration: KIŠIB šu-(d)EN.LÍL DUMU šu-ku-bi-im KIŠIB ṣí-lu-lu DUMU ú-ku i-nu-mì i-dí-a-bu-um a-wa-sú iq-bi-ú 10 ma-na KÙ.BABBAR a-na ša-lim-a-šùr i-dí-in um-ma šu-ut-ma i-ṣí-ba-at KÙ-pì-a li-il₅-qé
\nTranslation: Seal of Šu-Illil son of Šu-Kūbum, seal of Ṣilūlu son of Uku. When Iddin-abum spoke his will, he gave 10 minas ofדsilver to Šalim-Aššur. He said: He may take it from the interest on my silver.\"\"
\nExample
\nTransliteration: um-ma šu-ku-tum-ma a-na IŠTAR-lá-ma-sí ù ni-ta-aḫ-šu-šar qí-bi₄-ma mì-šu ša ta-áš-pu-ra-ni-ni um-ma a-tí-na-ma É-tum a-na lá be-tim i-tù-ar a-pu-tum a-na en-um-a-šùr i-xx-ni-ma e ší-na ga x ša lá ta-ḫa-dì-ri a-na IŠTAR-lá-ma-sí qí-bi₄-ma šu-ma a-ḫa-tí a-ta li-ba-am dì-ni-ší-im lá ta-ḫa-da-ar a-na ni-ta-aḫ-šu-šar qí-bi₄-ma TÚG-pì-ri-kà-ni ša e-zi-bu na-pí-ší-šu-nu ù ṭup-pu-ú lu ša-ṣú-ru pì-ri-kà-nu ša ma-tí ù tí-bu-lá ma-a-x iš-ta-ú-mu-ni a-dí en-um-a-šùr i-lá-kà-ni a ma-ma-an lá tù-šé-ri x GÍN KÙ.BABBAR (d)UTU-tap-pá-i ub-lá-ki-im 1 GÍN KÙ.GI ù x GÍN KÙ.BABBAR i-ku-pì-a ub-lá-ki-im
\nTranslation: From Šukkutum to Ištar-lamassī and Nitahšušar: Why is that you (fem. plur.) have written me, saying: The house is no longer a house.\" Urgent, to Ennam-Aššur … Do not fear!. To Ištar-lamassī: If you are truly my sister, then encourage her. Do not fear. To Nitahšušar: Air the -textiles that I left. Also, the tablets should be guarded. The which Mati? and Tibula … have bought do not release (them) to anyone before Ennam-Aššur arrives. Šamaš-tappā'ī brought you x shekels of silver. Ikūn-pīya brought you 1 shekel of gold and x shekels of silver. \"
\nExample
\nTransliteration: 30 ma-na KÙ.BABBAR ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a-na um-mì-a-ni-a ša ki-ma i-a-tí ù ḫi-na-a šu-ku-bu-um ú-bi₄-il₅ IGI a-mur-(d)UTU DUMU mì-šar-ra-bí IGI a-mur-(d)UTU DUMU ḫa-ri-im-tim 2 ma-na KÙ.BABBAR i-a-am 1 ma-na KÙ.BABBAR ša ḫi-na-a a pá-ni ILLAT-tim 1.5 ma-na KÙ.BABBAR ša um-mì-iš-ḫa-ra 15 GÍN KÙ.KI ša lá-ma-sí-tim ú a-šur-iš-tí-kál ŠÀ.BA 0.5 ma-na 5 GÍN ša lá-ma-sí-tim 0.33333 ma-na 5 GÍN KÙ.BABBAR ša ik-ri-bi-a mì-ma a-nim ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a ša ki-ma i-a-tí šu-ku-bu-um ú-bi₄-il₅ IGI lá-qé-ep DUMU i-zi-za-am-ì-lí IGI a-šur-na-da
\nTranslation: 30 minas of silver, its excise [added], his transport fee paid, under my seals, Šu-Kūbum brought to my investors and my representatives and to hinnāya. (Entrusted to him) in the presence of Amur-Šamaš, son of Mīšar-rabi, of Amur-Šamaš, son of harimtum. 2 minas of silver belonging to me; 1 mina of silver belonging to hinnāya, (sent) to meet the caravan; 1.5 mina of silver of Ummī-Išhara; 15 shekels of gold of Lamassutum and Aššuriš-tikal, thereof 35 shekels of silver of Lamassutum (and?) 25 shekels of [silver] being [my?] temple funds - all this, its import tax [added], his transport fee [paid], under my seals, Šu-Kūbum brought to my representatives. In the presence of Lā-qēp, son of Izzizam-ilī, of Aššur-nādā.
\nExample (it's the fractions)
\nTransliteration: 2 né-pí-šu 15 ma-na.TA ú iš-tí-in né-pí-šu-um 10 ma-na ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu-ú ŠU.NÍGIN 42.33333 ma-na KÙ.BABBAR ṣa-ru-pá-am ku-nu-ki-a a-na a-lu-wa ù e-ni-ša-ri-im áp-qí-id-ma a-na a-lim(ki) a ma-lá tí-ir-tí-šu a-ṣé-er ša-lim-a-šùr ú-šé-bi-il₅-šu-nu a-ḫa-ma 13.33333 ma-na URUDU SIG₅ a-na ga-am-ri-šu-nu ù 5 GÍN KÙ.BABBAR a-na ú-ku-ul-tí-šu-nu a-dí-in IGI ili₅-ba-ni DUMU ba-ší-lam IGI a-ḫu-qar DUMU zu-ur-zu-ur IGI tù-ra-am-ì-lí DUMU e-dí-na-a-šùr a-ḫa-ma 10.33333 ma-na 3.5 GÍN KÙ.BABBAR ṣa-ru-pá-am x tum ni x x x 0.5 GÍN ṣí-ba-tim ša i-na ṣé-ri-a il₅-qé-ú-ni a-na ḫu-bu-li-šu a-na kà-ri-im wa-aḫ-šu-ša-na áš-qúl ṣí-ba-at KÙ.GI ša ší-ip-kà-at a-šur-bé-el-a-wa-tim i-ší-tù
\nTranslation: 2 packages of 15 minas each plus a single package of 10 minas, its import duty added, its transport tariff paid - in all 42 ⅓ minas of refined silver under my seal, I entrusted to Aluwa and Enišārum, and I sent them to the City to Šalim-Aššur in accordance with his orders. Furthermore, I paid 13 ⅓ minas of good copper for their expenses and 5 shekels of silver for their food. Witnessed by Ilī-bāni son of Baši-ilum, by Ahu-qar son of Zurzur, by Tūram-ilī son of Eddin-Aššur. Furthermore, 10 ⅓ minas 3 ½ shekels of refined silver …, … ½, shekel interest that they took on my account, I paid for his debt to the Wahšušana colony. The interest on the gold that remained of Aššur-bēl-awātim's investment.
\nExample
\nTransliteration: 4 GÍN a-na ší-iṭ-ri-im ša pu-ki-im 1.3333300000000001 GÍN a-na e-re-qí-im qá-nu-e áš-qúl 1.66666 GÍN a sú-ba-ri-im áš-qúl 0.33333 ma-na KÙ.BABBAR a-na a-bar-ni-im áš-qúl 0.16666 GÍN a-na um-ṣí-im 0.25 GÍN a šu-um-ki xxx x na-ru-uq GIG xxxx […] xx GÍN a-na … ḫa-áš-lá-tim xxx a-wa-ar-nu-a-lim 0.66666 GÍN a-na e-ṣé áš-qúl 0.66666 GÍN a pá-e ú-ša-qí-il₅ 1 bi-il₅-té-en 0.5 GÍN áš-qúl 0.25 GÍN a-na na-pá-ḫi-im 95 ki-ra-tim a-na x 0.25 GÍN.TA ù 7.5 ŠE.TA 0.5 GÍN a-na xxxx 0.16666 GÍN a-na xxxx a-na xxxx
\nTranslation: I paid 4 shekels for a scarf of -weave, 1.3333 shekel for a wagonload of reed; I paid 1.6666 shekel for ; I paid 0.3333 mina of silver for an Abarnian textile, 1 / 6 shekel for a piece of dried meat, 1 / 4 shekel for onions, [x shekels for] [x] bags of wheat… …;[I paid … … … … … …] shek[els for] ?, [… … … …] for a … … … …; I paid 0.6666 shekel for firewood; 0.6666 shekels I paid for chaff, for a double load I paid 0.5 shekel; 1 / 4 shekel for the blacksmith. I supplied 95 drinks to .. at 52.5 grains (of silver) each, 0.5 shekel for … … … 1 / 6 shekel for … for … … … … … … …
\nExample
\nTransliteration: 1.25 GÍN KÙ.BABBAR i-na li-bi₄ lu-lu
\nTranslation: 1 1 / 4 shekel of silver is owed by Lulu.
\nExample
\nTransliteration: […] x šu-a-x-… um-ma … … … … en-nam-a-šur … bi-ma … tám-kà-ri-im … ḫu-a … … KÙ.BABBAR IGI e-ni-iš-ta-ru-um … ni AN.NA-ki … ší-a-ni … … ma x ku-un-kà-ma a-na e-ni-iš-ta-ri-im dí-na-ma iš-tí ba-tí-qí-im ša a-na pá-ni ILLAT-tim i-lá-kà-ni lá i-sà-ḫu-ur ṭù-ur-da-ni-šu a-na-kam we-da-ku a-ḫu-a a-tù-nu e-ni-iš-ta-ra-am lá i-sà-ḫu-ur ṭù-ur-da-ni-šu mì-li-ik lu-qú-tim x x bi-ri-im mì-il₅-kà sà-ḫe-er-tám ša 5 GÍN KÙ.BABBAR x x nim a-na ší-im … … … … mì-li-ik-šu-nu mì-il₅-kà
\nTranslation: [Obverse too broken for translation]
\nExample
\nTransliteration: lu-lu en-nam-a-šur ù a-bu-um-ša-lim ša ki-ma a-šùr-DU₁₀ DUMU e-lá-ma a-na a-šur-DU₁₀ DUMU a-lá-ḫi-im iṣ-bu-tù-ni-a-tí-ma um-ma šu-nu-ma ší-im 21 GÚ 9 ma-na URUDU ma-sí-im ša a-šur-DU₁₀ DUMU e-lá-ma ša i-na É-tí-kà i-ni-id-ú a-na ší-mì-im ta-dí-šu um-ma a-šur-DU₁₀-ma a-dí-šu-ma 17.83333 ma-na 2 GÍN KÙ.BABBAR ší-im-šu na-áš-a-ku eq-lam a šu-mì a-šur-DU₁₀ DUMU e-lá-ma KÙ.BABBAR e-tí-iq a-na a-wa-tim a-ni-a-tim kà-ru-um bu-ru-uš-ḫa-tum i-dí-ni-a-tí-ma ší-bu-tí-ni IGI GÍR ša a-šur ni-dí-in IGI i-ku-pì-a IGI ú-ṣú-ur-ša-a-šur
\nTranslation: Lullu,
\nThere's a lot more, like transliterations with lots of gaps but no '…' in the translation or the opposite scenario. I may share a notebook that contains more examples if there's any interest. Any help would be appreciated.
", "rawMarkdown": "After working on cleaning the text for modelling, I've found some transliteration/translation pairs that seem weird and was hoping to get some clarification on if this stuff will be seen in the hidden test set or if it has been cleaned in a certain way that we should replicate. \n\nI'm bolding the text I'm questioning in the following examples. These examples were not cleaned.\n\n**Example**\n\nTransliteration: KIŠIB šu-(d)EN.LÍL DUMU šu-ku-bi-im KIŠIB ṣí-lu-lu DUMU ú-ku i-nu-mì i-dí-a-bu-um a-wa-sú iq-bi-ú 10 ma-na KÙ.BABBAR a-na ša-lim-a-šùr i-dí-in um-ma šu-ut-ma i-ṣí-ba-at KÙ-pì-a li-il₅-qé\n\nTranslation: Seal of Šu-Illil son of Šu-Kūbum, seal of Ṣilūlu son of Uku. When Iddin-abum spoke his will, he gave 10 minas **ofדsilver** to Šalim-Aššur. He said: He may take it from the interest on my silver.\"\"\n\n**Example**\n\nTransliteration: um-ma šu-ku-tum-ma a-na IŠTAR-lá-ma-sí ù ni-ta-aḫ-šu-šar qí-bi₄-ma mì-šu ša ta-áš-pu-ra-ni-ni um-ma a-tí-na-ma É-tum a-na lá be-tim i-tù-ar a-pu-tum a-na en-um-a-šùr i-xx-ni-ma e ší-na ga x ša lá ta-ḫa-dì-ri a-na IŠTAR-lá-ma-sí qí-bi₄-ma šu-ma a-ḫa-tí a-ta li-ba-am dì-ni-ší-im lá ta-ḫa-da-ar a-na ni-ta-aḫ-šu-šar qí-bi₄-ma TÚG-pì-ri-kà-ni ša e-zi-bu na-pí-ší-šu-nu ù ṭup-pu-ú lu ša-ṣú-ru pì-ri-kà-nu ša ma-tí ù tí-bu-lá ma-a-x iš-ta-ú-mu-ni a-dí en-um-a-šùr i-lá-kà-ni a ma-ma-an lá tù-šé-ri x GÍN KÙ.BABBAR (d)UTU-tap-pá-i ub-lá-ki-im 1 GÍN KÙ.GI ù x GÍN KÙ.BABBAR i-ku-pì-a ub-lá-ki-im\n\nTranslation: From Šukkutum to Ištar-lamassī and Nitahšušar: Why is that you **(fem. plur.)** have written me, saying: The house is no longer a house.\" Urgent, to Ennam-Aššur ... Do not fear!. To Ištar-lamassī: If you are truly my sister, then encourage her. Do not fear. To Nitahšušar: Air the -textiles that I left. Also, the tablets should be guarded. The which Mati? and Tibula ... have bought do not release **(them)** to anyone before Ennam-Aššur arrives. Šamaš-tappā'ī brought you x shekels of silver. Ikūn-pīya brought you 1 shekel of gold and x shekels of silver. \"\n\n**Example**\n\nTransliteration: 30 ma-na KÙ.BABBAR ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a-na um-mì-a-ni-a ša ki-ma i-a-tí ù ḫi-na-a šu-ku-bu-um ú-bi₄-il₅ IGI a-mur-(d)UTU DUMU mì-šar-ra-bí IGI a-mur-(d)UTU DUMU ḫa-ri-im-tim 2 ma-na KÙ.BABBAR i-a-am 1 ma-na KÙ.BABBAR ša ḫi-na-a a pá-ni ILLAT-tim 1.5 ma-na KÙ.BABBAR ša um-mì-iš-ḫa-ra 15 GÍN KÙ.KI ša lá-ma-sí-tim ú a-šur-iš-tí-kál ŠÀ.BA 0.5 ma-na 5 GÍN ša lá-ma-sí-tim 0.33333 ma-na 5 GÍN KÙ.BABBAR ša ik-ri-bi-a mì-ma a-nim ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a ša ki-ma i-a-tí šu-ku-bu-um ú-bi₄-il₅ IGI lá-qé-ep DUMU i-zi-za-am-ì-lí IGI a-šur-na-da\n\nTranslation: 30 minas of silver, its excise **[added]**, his transport fee paid, under my seals, Šu-Kūbum brought to my investors and my representatives and to hinnāya. **(Entrusted to him)** in the presence of Amur-Šamaš, son of Mīšar-rabi, of Amur-Šamaš, son of harimtum. 2 minas of silver belonging to me; 1 mina of silver belonging to hinnāya, **(sent)** to meet the caravan; 1.5 mina of silver of Ummī-Išhara; 15 shekels of gold of Lamassutum and Aššuriš-tikal, thereof 35 shekels of silver of Lamassutum **(and?)** 25 shekels of **[silver]** being **[my?]** temple funds - all this, its import tax **[added]**, his transport fee **[paid]**, under my seals, Šu-Kūbum brought to my representatives. In the presence of Lā-qēp, son of Izzizam-ilī, of Aššur-nādā.\n\n**Example** (it's the fractions)\n\nTransliteration: 2 né-pí-šu 15 ma-na.TA ú iš-tí-in né-pí-šu-um 10 ma-na ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu-ú ŠU.NÍGIN 42.33333 ma-na KÙ.BABBAR ṣa-ru-pá-am ku-nu-ki-a a-na a-lu-wa ù e-ni-ša-ri-im áp-qí-id-ma a-na a-lim(ki) a ma-lá tí-ir-tí-šu a-ṣé-er ša-lim-a-šùr ú-šé-bi-il₅-šu-nu a-ḫa-ma 13.33333 ma-na URUDU SIG₅ a-na ga-am-ri-šu-nu ù 5 GÍN KÙ.BABBAR a-na ú-ku-ul-tí-šu-nu a-dí-in IGI ili₅-ba-ni DUMU ba-ší-lam IGI a-ḫu-qar DUMU zu-ur-zu-ur IGI tù-ra-am-ì-lí DUMU e-dí-na-a-šùr a-ḫa-ma 10.33333 ma-na 3.5 GÍN KÙ.BABBAR ṣa-ru-pá-am x tum ni x x x 0.5 GÍN ṣí-ba-tim ša i-na ṣé-ri-a il₅-qé-ú-ni a-na ḫu-bu-li-šu a-na kà-ri-im wa-aḫ-šu-ša-na áš-qúl ṣí-ba-at KÙ.GI ša ší-ip-kà-at a-šur-bé-el-a-wa-tim i-ší-tù\n\nTranslation: 2 packages of 15 minas each plus a single package of 10 minas, its import duty added, its transport tariff paid - in all 42 **⅓** minas of refined silver under my seal, I entrusted to Aluwa and Enišārum, and I sent them to the City to Šalim-Aššur in accordance with his orders. Furthermore, I paid 13 **⅓** minas of good copper for their expenses and 5 shekels of silver for their food. Witnessed by Ilī-bāni son of Baši-ilum, by Ahu-qar son of Zurzur, by Tūram-ilī son of Eddin-Aššur. Furthermore, 10 **⅓** minas 3 **½** shekels of refined silver ..., ... **½**, shekel interest that they took on my account, I paid for his debt to the Wahšušana colony. The interest on the gold that remained of Aššur-bēl-awātim's investment.\n\n**Example**\n\nTransliteration: 4 GÍN a-na ší-iṭ-ri-im ša pu-ki-im **1.3333300000000001** GÍN a-na e-re-qí-im qá-nu-e áš-qúl 1.66666 GÍN a sú-ba-ri-im áš-qúl 0.33333 ma-na KÙ.BABBAR a-na a-bar-ni-im áš-qúl 0.16666 GÍN a-na um-ṣí-im 0.25 GÍN a šu-um-ki xxx x na-ru-uq GIG **xxxx** [...] xx GÍN a-na … ḫa-áš-lá-tim xxx a-wa-ar-nu-a-lim 0.66666 GÍN a-na e-ṣé áš-qúl 0.66666 GÍN a pá-e ú-ša-qí-il₅ 1 bi-il₅-té-en 0.5 GÍN áš-qúl 0.25 GÍN a-na na-pá-ḫi-im 95 ki-ra-tim a-na x 0.25 GÍN.TA ù 7.5 ŠE.TA 0.5 GÍN a-na **xxxx** 0.16666 GÍN a-na **xxxx** a-na xxxx\n\nTranslation: I paid 4 shekels for a scarf of -weave, 1.3333 shekel for a wagonload of reed; I paid 1.6666 shekel for ; I paid 0.3333 mina of silver for an Abarnian textile, **1 / 6** shekel for a piece of dried meat, **1 / 4** shekel for onions, **[x shekels for]** [x] bags of wheat... ...;**[I paid ... ... ... ... ... ...]** **shek[els for] ?**, [... ... ... ...] for a ... ... ... ...; I paid 0.6666 shekel for firewood; 0.6666 shekels I paid for chaff, for a double load I paid 0.5 shekel; **1 / 4** shekel for the blacksmith. I supplied 95 drinks to .. at 52.5 grains (of silver) each, 0.5 shekel for ... ... ... **1 / 6** shekel for ... for ... ... ... ... ... ... ...\n\n**Example**\n\nTransliteration: 1.25 GÍN KÙ.BABBAR i-na li-bi₄ lu-lu\n\nTranslation: **1 1 / 4** shekel of silver is owed by Lulu.\n\n**Example**\n\nTransliteration: [...] x šu-a-x-… um-ma … … … … en-nam-a-šur … bi-ma … tám-kà-ri-im … ḫu-a … … KÙ.BABBAR IGI e-ni-iš-ta-ru-um … ni AN.NA-ki … ší-a-ni … … ma x ku-un-kà-ma a-na e-ni-iš-ta-ri-im dí-na-ma iš-tí ba-tí-qí-im ša a-na pá-ni ILLAT-tim i-lá-kà-ni lá i-sà-ḫu-ur ṭù-ur-da-ni-šu a-na-kam we-da-ku a-ḫu-a a-tù-nu e-ni-iš-ta-ra-am lá i-sà-ḫu-ur ṭù-ur-da-ni-šu mì-li-ik lu-qú-tim x x bi-ri-im mì-il₅-kà sà-ḫe-er-tám ša 5 GÍN KÙ.BABBAR x x nim a-na ší-im … … … … mì-li-ik-šu-nu mì-il₅-kà\n\nTranslation: **[Obverse too broken for translation]**\n\n**Example**\n\nTransliteration: lu-lu en-nam-a-šur ù a-bu-um-ša-lim ša ki-ma a-šùr-DU₁₀ DUMU e-lá-ma a-na a-šur-DU₁₀ DUMU a-lá-ḫi-im iṣ-bu-tù-ni-a-tí-ma um-ma šu-nu-ma ší-im 21 GÚ 9 ma-na URUDU ma-sí-im ša a-šur-DU₁₀ DUMU e-lá-ma ša i-na É-tí-kà i-ni-id-ú a-na ší-mì-im ta-dí-šu um-ma a-šur-DU₁₀-ma a-dí-šu-ma 17.83333 ma-na 2 GÍN KÙ.BABBAR ší-im-šu na-áš-a-ku eq-lam a šu-mì a-šur-DU₁₀ DUMU e-lá-ma KÙ.BABBAR e-tí-iq a-na a-wa-tim a-ni-a-tim kà-ru-um bu-ru-uš-ḫa-tum i-dí-ni-a-tí-ma ší-bu-tí-ni IGI GÍR ša a-šur ni-dí-in IGI i-ku-pì-a IGI ú-ṣú-ur-ša-a-šur\n\nTranslation: Lullu,\n\nThere's a lot more, like transliterations with lots of gaps but no '...' in the translation or the opposite scenario. I may share a notebook that contains more examples if there's any interest. Any help would be appreciated.", "votes": null }, { "id": "3382662", "postDate": "12/28/2025 07:56:53", "content": "What are those? @jackvd
", "rawMarkdown": "What are those? @jackvd", "votes": null }, { "id": "3382916", "postDate": "12/28/2025 20:07:57", "content": "Thank you for your comment Jack! These errors and oversights in the data stem from the nature of the databases we used to gather enough texts for this challenge. Many of these databases are worked on by students with little oversight. That goes for the CDLI database as well. As mentioned elsewhere, English translations are rarely paired with transliterations in these databases. So what we have provided is the first attempt for Old Assyrian, but by no means an error-free dataset.
\nRegarding the specifics you mention:
\nExample 1. The parentheses used as determinatives (e.g. (d)UTU) should be replaced with curly brackets (e.g. {d}UTU). This came from a different dataset which we attempted to harmonize with the rest, but this was one difference which we missed.
\nExample 2. The first example includes scribal insertions in parentheses (fem. plur.). We have made sure to remove these from the hidden evaluation translations. So it would be best to remove the modern scribal insertions when you encounter them.
\nExample 3. For instances of bracketed insertions like, \"its excise [added]\", it is sufficient to simply remove the brackets, since the specific phraseology is what we're hoping to train in a MT model, rather than a strict adherence to which words in the text is broken for this challenge.
\nExample 4 & 6. The fractions are retained in the hidden test dataset. There you will find 2 ½ minas = 2 ½ ma-na, and NOT 2.5 ma-na. I realize this is a problem and I'm working on providing a supplemental dataset of texts with integer (and fraction) numbers, rather than floats. I also made sure to use the same single character fractions for ⅓ ⅔ etc. But there is a larger issue at hand in aligning these numerical values, and that is sometimes the translators do some math to convert the fractions to whole numbers in shekels, instead of a literal translation of minas and shekels (e.g. \"3 ⅔ ma-na ⅓ GÍN KÙ.B.\"--> \"23 minas 40 ⅓ shekels of silver\"). The bad news is that this is going to pose a challenge for ML in working with the number systems in use at that time. The good news is this will likely negatively impact everyone's score equally.
\nExample 5. This is a very broken (and messy text). The aspect you highlighted has an additional closed square bracket due to one of the ellipses having a closed bracket I believe (i.e. [I paid … … … … … …] shek[els for] ?). This can be remedied by using the substitutions described in the data, i.e. … = , and remove the square brackets (retaining the English translations).
\nExample 7 & 8. Texts like this should probably have been removed, an oversight on my part. I look forward to working together to get the best \"gold standard\" ML dataset for these texts!
", "rawMarkdown": "Thank you for your comment Jack! These errors and oversights in the data stem from the nature of the databases we used to gather enough texts for this challenge. Many of these databases are worked on by students with little oversight. That goes for the CDLI database as well. As mentioned elsewhere, English translations are rarely paired with transliterations in these databases. So what we have provided is the first attempt for Old Assyrian, but by no means an error-free dataset. \n\nRegarding the specifics you mention:\n\nExample 1. The parentheses used as determinatives (e.g. (d)UTU) should be replaced with curly brackets (e.g. {d}UTU). This came from a different dataset which we attempted to harmonize with the rest, but this was one difference which we missed.\n\nExample 2. The first example includes scribal insertions in parentheses (fem. plur.). We have made sure to remove these from the hidden evaluation translations. So it would be best to remove the modern scribal insertions when you encounter them.\n\nExample 3. For instances of bracketed insertions like, \"its excise [added]\", it is sufficient to simply remove the brackets, since the specific phraseology is what we're hoping to train in a MT model, rather than a strict adherence to which words in the text is broken for this challenge. \n\nExample 4 & 6. The fractions are retained in the hidden test dataset. There you will find 2 ½ minas = 2 ½ ma-na, and NOT 2.5 ma-na. I realize this is a problem and I'm working on providing a supplemental dataset of texts with integer (and fraction) numbers, rather than floats. I also made sure to use the same single character fractions for ⅓ ⅔ etc. But there is a larger issue at hand in aligning these numerical values, and that is sometimes the translators do some math to convert the fractions to whole numbers in shekels, instead of a literal translation of minas and shekels (e.g. \"3 ⅔ ma-na ⅓ GÍN KÙ.B.\"--> \"23 minas 40 ⅓ shekels of silver\"). The bad news is that this is going to pose a challenge for ML in working with the number systems in use at that time. The good news is this will likely negatively impact everyone's score equally.\n\nExample 5. This is a very broken (and messy text). The aspect you highlighted has an additional closed square bracket due to one of the ellipses having a closed bracket I believe (i.e. [I paid … … … … … …] shek[els for] ?). This can be remedied by using the substitutions described in the data, i.e. ... =Thanks! This is very helpful.
", "rawMarkdown": "Thanks! This is very helpful.", "votes": null }, { "id": "3383162", "postDate": "12/29/2025 14:51:28", "content": "Apologies, but here are some more little snippets that I'd like to confirm should be removed without concern: \"ofדsilver\",\"myself()\",\"< lil >\", \"< of firewood >\"
", "rawMarkdown": "Apologies, but here are some more little snippets that I'd like to confirm should be removed without concern: \"ofדsilver\",\"myself()\",\"< lil >\", \"< of firewood >\"", "votes": null }, { "id": "3383194", "postDate": "12/29/2025 16:05:04", "content": "yes, I have no idea how a Hebrew dalet ד got in there!
\n\"ofדsilver\" --> of silver\n\"myself()\" --> myself\n\"< lil >\" --> -lil — I'm assuming it is the end of a name like En-lil : Note that modern scribes / editors use these pointy brackets to insert signs which they think are missing in the text (i.e. ancient scribal omission).\n\"< of firewood >\" --> of firewood
", "rawMarkdown": "yes, I have no idea how a Hebrew dalet ד got in there! \n\n\"ofדsilver\" --> of silver\n\"myself()\" --> myself\n\"< lil >\" --> -lil — I'm assuming it is the end of a name like En-lil : Note that modern scribes / editors use these pointy brackets to insert signs which they think are missing in the text (i.e. ancient scribal omission).\n\"< of firewood >\" --> of firewood", "votes": null }, { "id": "3383335", "postDate": "12/29/2025 22:41:44", "content": "@deeppast How will you handle 1 / 12, 7 / 12, 11 / 12?
\nFor example:
\nLet me put it this way, there are no instances of '/' in either the transliterations or translations of the public or private evaluation data.
", "rawMarkdown": "Let me put it this way, there are no instances of '/' in either the transliterations or translations of the public or private evaluation data.", "votes": null } ], "comments": [ { "id": 3382662, "author_name": "navneetbende", "author_url": "", "post_date": "12/28/2025 07:56:53", "content": "What are those? @jackvd
", "votes": null, "replies": [] }, { "id": 3382916, "author_name": "deeppast", "author_url": "", "post_date": "12/28/2025 20:07:57", "content": "Thank you for your comment Jack! These errors and oversights in the data stem from the nature of the databases we used to gather enough texts for this challenge. Many of these databases are worked on by students with little oversight. That goes for the CDLI database as well. As mentioned elsewhere, English translations are rarely paired with transliterations in these databases. So what we have provided is the first attempt for Old Assyrian, but by no means an error-free dataset.
\nRegarding the specifics you mention:
\nExample 1. The parentheses used as determinatives (e.g. (d)UTU) should be replaced with curly brackets (e.g. {d}UTU). This came from a different dataset which we attempted to harmonize with the rest, but this was one difference which we missed.
\nExample 2. The first example includes scribal insertions in parentheses (fem. plur.). We have made sure to remove these from the hidden evaluation translations. So it would be best to remove the modern scribal insertions when you encounter them.
\nExample 3. For instances of bracketed insertions like, \"its excise [added]\", it is sufficient to simply remove the brackets, since the specific phraseology is what we're hoping to train in a MT model, rather than a strict adherence to which words in the text is broken for this challenge.
\nExample 4 & 6. The fractions are retained in the hidden test dataset. There you will find 2 ½ minas = 2 ½ ma-na, and NOT 2.5 ma-na. I realize this is a problem and I'm working on providing a supplemental dataset of texts with integer (and fraction) numbers, rather than floats. I also made sure to use the same single character fractions for ⅓ ⅔ etc. But there is a larger issue at hand in aligning these numerical values, and that is sometimes the translators do some math to convert the fractions to whole numbers in shekels, instead of a literal translation of minas and shekels (e.g. \"3 ⅔ ma-na ⅓ GÍN KÙ.B.\"--> \"23 minas 40 ⅓ shekels of silver\"). The bad news is that this is going to pose a challenge for ML in working with the number systems in use at that time. The good news is this will likely negatively impact everyone's score equally.
\nExample 5. This is a very broken (and messy text). The aspect you highlighted has an additional closed square bracket due to one of the ellipses having a closed bracket I believe (i.e. [I paid … … … … … …] shek[els for] ?). This can be remedied by using the substitutions described in the data, i.e. … = , and remove the square brackets (retaining the English translations).
\nExample 7 & 8. Texts like this should probably have been removed, an oversight on my part. I look forward to working together to get the best \"gold standard\" ML dataset for these texts!
", "votes": null, "replies": [ { "id": 3382950, "author_name": "jackvd", "author_url": "", "post_date": "12/28/2025 22:14:06", "content": "Thanks! This is very helpful.
", "votes": null, "replies": [] }, { "id": 3383162, "author_name": "jackvd", "author_url": "", "post_date": "12/29/2025 14:51:28", "content": "Apologies, but here are some more little snippets that I'd like to confirm should be removed without concern: \"ofדsilver\",\"myself()\",\"< lil >\", \"< of firewood >\"
", "votes": null, "replies": [ { "id": 3383194, "author_name": "deeppast", "author_url": "", "post_date": "12/29/2025 16:05:04", "content": "yes, I have no idea how a Hebrew dalet ד got in there!
\n\"ofדsilver\" --> of silver\n\"myself()\" --> myself\n\"< lil >\" --> -lil — I'm assuming it is the end of a name like En-lil : Note that modern scribes / editors use these pointy brackets to insert signs which they think are missing in the text (i.e. ancient scribal omission).\n\"< of firewood >\" --> of firewood
", "votes": null, "replies": [] } ] }, { "id": 3383335, "author_name": "mpware", "author_url": "", "post_date": "12/29/2025 22:41:44", "content": "@deeppast How will you handle 1 / 12, 7 / 12, 11 / 12?
\nFor example:
\nLet me put it this way, there are no instances of '/' in either the transliterations or translations of the public or private evaluation data.
", "votes": null, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "3382583": "After working on cleaning the text for modelling, I've found some transliteration/translation pairs that seem weird and was hoping to get some clarification on if this stuff will be seen in the hidden test set or if it has been cleaned in a certain way that we should replicate. \n\nI'm bolding the text I'm questioning in the following examples. These examples were not cleaned.\n\n**Example**\n\nTransliteration: KIŠIB šu-(d)EN.LÍL DUMU šu-ku-bi-im KIŠIB ṣí-lu-lu DUMU ú-ku i-nu-mì i-dí-a-bu-um a-wa-sú iq-bi-ú 10 ma-na KÙ.BABBAR a-na ša-lim-a-šùr i-dí-in um-ma šu-ut-ma i-ṣí-ba-at KÙ-pì-a li-il₅-qé\n\nTranslation: Seal of Šu-Illil son of Šu-Kūbum, seal of Ṣilūlu son of Uku. When Iddin-abum spoke his will, he gave 10 minas **ofדsilver** to Šalim-Aššur. He said: He may take it from the interest on my silver.\"\"\n\n**Example**\n\nTransliteration: um-ma šu-ku-tum-ma a-na IŠTAR-lá-ma-sí ù ni-ta-aḫ-šu-šar qí-bi₄-ma mì-šu ša ta-áš-pu-ra-ni-ni um-ma a-tí-na-ma É-tum a-na lá be-tim i-tù-ar a-pu-tum a-na en-um-a-šùr i-xx-ni-ma e ší-na ga x ša lá ta-ḫa-dì-ri a-na IŠTAR-lá-ma-sí qí-bi₄-ma šu-ma a-ḫa-tí a-ta li-ba-am dì-ni-ší-im lá ta-ḫa-da-ar a-na ni-ta-aḫ-šu-šar qí-bi₄-ma TÚG-pì-ri-kà-ni ša e-zi-bu na-pí-ší-šu-nu ù ṭup-pu-ú lu ša-ṣú-ru pì-ri-kà-nu ša ma-tí ù tí-bu-lá ma-a-x iš-ta-ú-mu-ni a-dí en-um-a-šùr i-lá-kà-ni a ma-ma-an lá tù-šé-ri x GÍN KÙ.BABBAR (d)UTU-tap-pá-i ub-lá-ki-im 1 GÍN KÙ.GI ù x GÍN KÙ.BABBAR i-ku-pì-a ub-lá-ki-im\n\nTranslation: From Šukkutum to Ištar-lamassī and Nitahšušar: Why is that you **(fem. plur.)** have written me, saying: The house is no longer a house.\" Urgent, to Ennam-Aššur ... Do not fear!. To Ištar-lamassī: If you are truly my sister, then encourage her. Do not fear. To Nitahšušar: Air the -textiles that I left. Also, the tablets should be guarded. The which Mati? and Tibula ... have bought do not release **(them)** to anyone before Ennam-Aššur arrives. Šamaš-tappā'ī brought you x shekels of silver. Ikūn-pīya brought you 1 shekel of gold and x shekels of silver. \"\n\n**Example**\n\nTransliteration: 30 ma-na KÙ.BABBAR ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a-na um-mì-a-ni-a ša ki-ma i-a-tí ù ḫi-na-a šu-ku-bu-um ú-bi₄-il₅ IGI a-mur-(d)UTU DUMU mì-šar-ra-bí IGI a-mur-(d)UTU DUMU ḫa-ri-im-tim 2 ma-na KÙ.BABBAR i-a-am 1 ma-na KÙ.BABBAR ša ḫi-na-a a pá-ni ILLAT-tim 1.5 ma-na KÙ.BABBAR ša um-mì-iš-ḫa-ra 15 GÍN KÙ.KI ša lá-ma-sí-tim ú a-šur-iš-tí-kál ŠÀ.BA 0.5 ma-na 5 GÍN ša lá-ma-sí-tim 0.33333 ma-na 5 GÍN KÙ.BABBAR ša ik-ri-bi-a mì-ma a-nim ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu ku-nu-ki-a a ša ki-ma i-a-tí šu-ku-bu-um ú-bi₄-il₅ IGI lá-qé-ep DUMU i-zi-za-am-ì-lí IGI a-šur-na-da\n\nTranslation: 30 minas of silver, its excise **[added]**, his transport fee paid, under my seals, Šu-Kūbum brought to my investors and my representatives and to hinnāya. **(Entrusted to him)** in the presence of Amur-Šamaš, son of Mīšar-rabi, of Amur-Šamaš, son of harimtum. 2 minas of silver belonging to me; 1 mina of silver belonging to hinnāya, **(sent)** to meet the caravan; 1.5 mina of silver of Ummī-Išhara; 15 shekels of gold of Lamassutum and Aššuriš-tikal, thereof 35 shekels of silver of Lamassutum **(and?)** 25 shekels of **[silver]** being **[my?]** temple funds - all this, its import tax **[added]**, his transport fee **[paid]**, under my seals, Šu-Kūbum brought to my representatives. In the presence of Lā-qēp, son of Izzizam-ilī, of Aššur-nādā.\n\n**Example** (it's the fractions)\n\nTransliteration: 2 né-pí-šu 15 ma-na.TA ú iš-tí-in né-pí-šu-um 10 ma-na ni-is-ḫa-sú DIRI ša-du-a-sú ša-bu-ú ŠU.NÍGIN 42.33333 ma-na KÙ.BABBAR ṣa-ru-pá-am ku-nu-ki-a a-na a-lu-wa ù e-ni-ša-ri-im áp-qí-id-ma a-na a-lim(ki) a ma-lá tí-ir-tí-šu a-ṣé-er ša-lim-a-šùr ú-šé-bi-il₅-šu-nu a-ḫa-ma 13.33333 ma-na URUDU SIG₅ a-na ga-am-ri-šu-nu ù 5 GÍN KÙ.BABBAR a-na ú-ku-ul-tí-šu-nu a-dí-in IGI ili₅-ba-ni DUMU ba-ší-lam IGI a-ḫu-qar DUMU zu-ur-zu-ur IGI tù-ra-am-ì-lí DUMU e-dí-na-a-šùr a-ḫa-ma 10.33333 ma-na 3.5 GÍN KÙ.BABBAR ṣa-ru-pá-am x tum ni x x x 0.5 GÍN ṣí-ba-tim ša i-na ṣé-ri-a il₅-qé-ú-ni a-na ḫu-bu-li-šu a-na kà-ri-im wa-aḫ-šu-ša-na áš-qúl ṣí-ba-at KÙ.GI ša ší-ip-kà-at a-šur-bé-el-a-wa-tim i-ší-tù\n\nTranslation: 2 packages of 15 minas each plus a single package of 10 minas, its import duty added, its transport tariff paid - in all 42 **⅓** minas of refined silver under my seal, I entrusted to Aluwa and Enišārum, and I sent them to the City to Šalim-Aššur in accordance with his orders. Furthermore, I paid 13 **⅓** minas of good copper for their expenses and 5 shekels of silver for their food. Witnessed by Ilī-bāni son of Baši-ilum, by Ahu-qar son of Zurzur, by Tūram-ilī son of Eddin-Aššur. Furthermore, 10 **⅓** minas 3 **½** shekels of refined silver ..., ... **½**, shekel interest that they took on my account, I paid for his debt to the Wahšušana colony. The interest on the gold that remained of Aššur-bēl-awātim's investment.\n\n**Example**\n\nTransliteration: 4 GÍN a-na ší-iṭ-ri-im ša pu-ki-im **1.3333300000000001** GÍN a-na e-re-qí-im qá-nu-e áš-qúl 1.66666 GÍN a sú-ba-ri-im áš-qúl 0.33333 ma-na KÙ.BABBAR a-na a-bar-ni-im áš-qúl 0.16666 GÍN a-na um-ṣí-im 0.25 GÍN a šu-um-ki xxx x na-ru-uq GIG **xxxx** [...] xx GÍN a-na … ḫa-áš-lá-tim xxx a-wa-ar-nu-a-lim 0.66666 GÍN a-na e-ṣé áš-qúl 0.66666 GÍN a pá-e ú-ša-qí-il₅ 1 bi-il₅-té-en 0.5 GÍN áš-qúl 0.25 GÍN a-na na-pá-ḫi-im 95 ki-ra-tim a-na x 0.25 GÍN.TA ù 7.5 ŠE.TA 0.5 GÍN a-na **xxxx** 0.16666 GÍN a-na **xxxx** a-na xxxx\n\nTranslation: I paid 4 shekels for a scarf of -weave, 1.3333 shekel for a wagonload of reed; I paid 1.6666 shekel for ; I paid 0.3333 mina of silver for an Abarnian textile, **1 / 6** shekel for a piece of dried meat, **1 / 4** shekel for onions, **[x shekels for]** [x] bags of wheat... ...;**[I paid ... ... ... ... ... ...]** **shek[els for] ?**, [... ... ... ...] for a ... ... ... ...; I paid 0.6666 shekel for firewood; 0.6666 shekels I paid for chaff, for a double load I paid 0.5 shekel; **1 / 4** shekel for the blacksmith. I supplied 95 drinks to .. at 52.5 grains (of silver) each, 0.5 shekel for ... ... ... **1 / 6** shekel for ... for ... ... ... ... ... ... ...\n\n**Example**\n\nTransliteration: 1.25 GÍN KÙ.BABBAR i-na li-bi₄ lu-lu\n\nTranslation: **1 1 / 4** shekel of silver is owed by Lulu.\n\n**Example**\n\nTransliteration: [...] x šu-a-x-… um-ma … … … … en-nam-a-šur … bi-ma … tám-kà-ri-im … ḫu-a … … KÙ.BABBAR IGI e-ni-iš-ta-ru-um … ni AN.NA-ki … ší-a-ni … … ma x ku-un-kà-ma a-na e-ni-iš-ta-ri-im dí-na-ma iš-tí ba-tí-qí-im ša a-na pá-ni ILLAT-tim i-lá-kà-ni lá i-sà-ḫu-ur ṭù-ur-da-ni-šu a-na-kam we-da-ku a-ḫu-a a-tù-nu e-ni-iš-ta-ra-am lá i-sà-ḫu-ur ṭù-ur-da-ni-šu mì-li-ik lu-qú-tim x x bi-ri-im mì-il₅-kà sà-ḫe-er-tám ša 5 GÍN KÙ.BABBAR x x nim a-na ší-im … … … … mì-li-ik-šu-nu mì-il₅-kà\n\nTranslation: **[Obverse too broken for translation]**\n\n**Example**\n\nTransliteration: lu-lu en-nam-a-šur ù a-bu-um-ša-lim ša ki-ma a-šùr-DU₁₀ DUMU e-lá-ma a-na a-šur-DU₁₀ DUMU a-lá-ḫi-im iṣ-bu-tù-ni-a-tí-ma um-ma šu-nu-ma ší-im 21 GÚ 9 ma-na URUDU ma-sí-im ša a-šur-DU₁₀ DUMU e-lá-ma ša i-na É-tí-kà i-ni-id-ú a-na ší-mì-im ta-dí-šu um-ma a-šur-DU₁₀-ma a-dí-šu-ma 17.83333 ma-na 2 GÍN KÙ.BABBAR ší-im-šu na-áš-a-ku eq-lam a šu-mì a-šur-DU₁₀ DUMU e-lá-ma KÙ.BABBAR e-tí-iq a-na a-wa-tim a-ni-a-tim kà-ru-um bu-ru-uš-ḫa-tum i-dí-ni-a-tí-ma ší-bu-tí-ni IGI GÍR ša a-šur ni-dí-in IGI i-ku-pì-a IGI ú-ṣú-ur-ša-a-šur\n\nTranslation: Lullu,\n\nThere's a lot more, like transliterations with lots of gaps but no '...' in the translation or the opposite scenario. I may share a notebook that contains more examples if there's any interest. Any help would be appreciated.", "3382662": "What are those? @jackvd", "3382916": "Thank you for your comment Jack! These errors and oversights in the data stem from the nature of the databases we used to gather enough texts for this challenge. Many of these databases are worked on by students with little oversight. That goes for the CDLI database as well. As mentioned elsewhere, English translations are rarely paired with transliterations in these databases. So what we have provided is the first attempt for Old Assyrian, but by no means an error-free dataset. \n\nRegarding the specifics you mention:\n\nExample 1. The parentheses used as determinatives (e.g. (d)UTU) should be replaced with curly brackets (e.g. {d}UTU). This came from a different dataset which we attempted to harmonize with the rest, but this was one difference which we missed.\n\nExample 2. The first example includes scribal insertions in parentheses (fem. plur.). We have made sure to remove these from the hidden evaluation translations. So it would be best to remove the modern scribal insertions when you encounter them.\n\nExample 3. For instances of bracketed insertions like, \"its excise [added]\", it is sufficient to simply remove the brackets, since the specific phraseology is what we're hoping to train in a MT model, rather than a strict adherence to which words in the text is broken for this challenge. \n\nExample 4 & 6. The fractions are retained in the hidden test dataset. There you will find 2 ½ minas = 2 ½ ma-na, and NOT 2.5 ma-na. I realize this is a problem and I'm working on providing a supplemental dataset of texts with integer (and fraction) numbers, rather than floats. I also made sure to use the same single character fractions for ⅓ ⅔ etc. But there is a larger issue at hand in aligning these numerical values, and that is sometimes the translators do some math to convert the fractions to whole numbers in shekels, instead of a literal translation of minas and shekels (e.g. \"3 ⅔ ma-na ⅓ GÍN KÙ.B.\"--> \"23 minas 40 ⅓ shekels of silver\"). The bad news is that this is going to pose a challenge for ML in working with the number systems in use at that time. The good news is this will likely negatively impact everyone's score equally.\n\nExample 5. This is a very broken (and messy text). The aspect you highlighted has an additional closed square bracket due to one of the ellipses having a closed bracket I believe (i.e. [I paid … … … … … …] shek[els for] ?). This can be remedied by using the substitutions described in the data, i.e. ... =