Apart from linguistic transformations, synthesizing among statements usually requires specific synthesis rulesofstatements.

However, statements different from the original statement could also be logically consistent with respecttoagivencontext.

However, there's a lack of insight into how these models perform in terms

In this work, we study the performance of byte-level and subword-level autoregressive models when trained using a fixed compute budget.

Transformers with linear attention (i.e., linear transfor mers) and state-space models have recently been suggested as a viable linear-time alt ernative to transformers with softmax attention. However, these models still underp erform transformers especially on tasks that require in-context retrieval. Whil e more expressive variants of linear transformers which replace the additive upda te in linear transformers with the delta rule [DeltaNet; 101 ] have been found to be more effective at associative recall, existing algorithms for training such mode ls do not parallelize over sequence length and are thus inefficient to train on modern ha rdware. This work describes a hardware-efficient algorithm for training line ar transformers with the delta rule, which exploits a memory-efficient representati on for computing products of Householder matrices [ 11 ]. This algorithm allows us to scale up DeltaNet to standard language modeling settings. We train a 1.3B mode l for 100B tokens and find that it outperforms recent linear-time baselines su ch as Mamba [ 31 ] and GLA [ 124 ] in terms of perplexity and zero-shot performance on downst ream tasks. We also experiment with two hybrid models which combine Delt aNet layers with (1) sliding-window attention layers every other layer or (2) two global attention layers, and find that these hybrids outperform strong transf ormer baselines.

Sign language translation (SL T) addresses the problem of translating information from a sign language in video to a spoken language in text. Existing studies, while showing progress, are often limited to narrow domains and/or few sign languages and struggle with open-domain tasks. In this paper, we push forward the frontier of SL T by scaling pretraining data, model size, and number of translation directions. We perform large-scale SL T pretraining on different data including 1) noisy multilingual Y ouTube SL T data, 2) parallel text corpora, and 3) SL T data augmented by translating video captions to other languages with off-the-shelf machine translation models. We unify different pretraining tasks with task-specific prompts under the encoder-decoder architecture, and initialize the SL T model with pretrained (m/By)T5 models across model sizes. SL T pretraining results on How2Sign and FLEURS-ASL#0 (ASL to 42 spoken languages) demonstrate the significance of data/model scaling and cross-lingual cross-modal transfer, as well as the feasibility of zero-shot SL T. We finetune the pretrained SL T models on 5 downstream open-domain SL T benchmarks covering 5 sign languages. Experiments show substantial quality improvements over the vanilla baselines, surpassing the previous state-of-the-art (SOT A) by wide margins.

To ensure large language models (LLMs) are used safely, one must reduce their propensity to hallucinate or to generate unacceptable answers. A simple and often used strategy is to first let the LLM generate multiple hypotheses and then employ a reranker to choose the best one.