REVIEW 5 cited by
Tower+: Bridging Generality and Translation Specialization in Multilingual LLMs
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Fine-tuning pretrained LLMs has been shown to be an effective strategy for reaching state-of-the-art performance on specific tasks like machine translation. However, this process of adaptation often implies sacrificing general-purpose capabilities, such as conversational reasoning and instruction-following, hampering the utility of the system in real-world applications that require a mixture of skills. In this paper, we introduce Tower+, a suite of models designed to deliver strong performance across both translation and multilingual general-purpose text capabilities. We achieve a Pareto frontier between translation specialization and multilingual general-purpose capabilities by introducing a novel training recipe that builds on Tower (Alves et al., 2024), comprising continued pretraining, supervised fine-tuning, preference optimization, and reinforcement learning with verifiable rewards. At each stage of training, we carefully generate and curate data to strengthen performance on translation as well as general-purpose tasks involving code generation, mathematics problem solving, and general instruction-following. We develop models at multiple scales: 2B, 9B, and 72B. Our smaller models often outperform larger general-purpose open-weight and proprietary LLMs (e.g., Llama 3.3 70B, GPT-4o). Our largest model delivers best-in-class translation performance for high-resource languages and top results in multilingual Arena Hard evaluations and in IF-MT, a benchmark we introduce for evaluating both translation and instruction-following. Our findings highlight that it is possible to rival frontier models in general capabilities, while optimizing for specific business domains, such as translation and localization.
Forward citations
Cited by 5 Pith papers
-
The Role of Disfluencies in Speech Translation
Current speech translation systems delete disfluencies, and false starts and self-repairs cause most translation-quality loss, as shown by a new eight-language benchmark.
-
Learning to Translate Ambiguous Terminology by Preference Optimization on Post-Edits
Masked preference optimization on human post-edits teaches an LLM translator to pick the contextually correct term among many valid dictionary options, improving term accuracy from 53.7% to 56.3% on English-German.
-
Scaling Model and Data for Multilingual Machine Translation with Open Large Language Models
Open 12B multilingual translator MiLMMT-46, trained with up to 3B tokens of parallel data per language plus ~100K curated instruction pairs, matches proprietary systems on FLORES+ and WMT24++.
-
$M^2PO$: Multi-Perspective Multi-Pair Preference Optimization for Machine Translation
M2PO combines a QE-plus-alignment reward with a dynamic curriculum and multi-pair DPO loss, and reports WMT21-22 gains for a 7B translation model, but the abstract's WMT23/24 9B parity claims are unsupported.
-
Hunyuan-MT Technical Report
Hunyuan-MT and Chimera, a 7B open-source translation model and its multi-candidate fusion variant, claim state-of-the-art multilingual translation including Mandarin to minority languages, with open weights.
Discussion (0). Sign in to comment.