pith. machine review for the scientific record. sign in

arxiv: 2305.10403 · v3 · submitted 2023-05-17 · 💻 cs.CL · cs.AI

Recognition: 2 theorem links

· Lean Theorem

PaLM 2 Technical Report

Rohan Anil , Andrew M. Dai , Orhan Firat , Melvin Johnson , Dmitry Lepikhin , Alexandre Passos , Siamak Shakeri , Emanuel Taropa
show 120 more authors
Paige Bailey Zhifeng Chen Eric Chu Jonathan H. Clark Laurent El Shafey Yanping Huang Kathy Meier-Hellstern Gaurav Mishra Erica Moreira Mark Omernick Kevin Robinson Sebastian Ruder Yi Tay Kefan Xiao Yuanzhong Xu Yujing Zhang Gustavo Hernandez Abrego Junwhan Ahn Jacob Austin Paul Barham Jan Botha James Bradbury Siddhartha Brahma Kevin Brooks Michele Catasta Yong Cheng Colin Cherry Christopher A. Choquette-Choo Aakanksha Chowdhery Cl\'ement Crepy Shachi Dave Mostafa Dehghani Sunipa Dev Jacob Devlin Mark D\'iaz Nan Du Ethan Dyer Vlad Feinberg Fangxiaoyu Feng Vlad Fienber Markus Freitag Xavier Garcia Sebastian Gehrmann Lucas Gonzalez Guy Gur-Ari Steven Hand Hadi Hashemi Le Hou Joshua Howland Andrea Hu Jeffrey Hui Jeremy Hurwitz Michael Isard Abe Ittycheriah Matthew Jagielski Wenhao Jia Kathleen Kenealy Maxim Krikun Sneha Kudugunta Chang Lan Katherine Lee Benjamin Lee Eric Li Music Li Wei Li Yaguang Li Jian Li Hyeontaek Lim Hanzhao Lin Zhongtao Liu Frederick Liu Marcello Maggioni Aroma Mahendru Joshua Maynez Vedant Misra Maysam Moussalem Zachary Nado John Nham Eric Ni Andrew Nystrom Alicia Parrish Marie Pellat Martin Polacek Alex Polozov Reiner Pope Siyuan Qiao Emily Reif Bryan Richter Parker Riley Alex Castro Ros Aurko Roy Brennan Saeta Rajkumar Samuel Renee Shelby Ambrose Slone Daniel Smilkov David R. So Daniel Sohn Simon Tokumine Dasha Valter Vijay Vasudevan Kiran Vodrahalli Xuezhi Wang Pidong Wang Zirui Wang Tao Wang John Wieting Yuhuai Wu Kelvin Xu Yunhan Xu Linting Xue Pengcheng Yin Jiahui Yu Qiao Zhang Steven Zheng Ce Zheng Weikang Zhou Denny Zhou Slav Petrov Yonghui Wu
Authors on Pith no claims yet

Pith reviewed 2026-05-12 11:54 UTC · model grok-4.3

classification 💻 cs.CL cs.AI
keywords PaLM 2language modelmultilingual capabilitiesreasoningcompute efficiencyTransformerbenchmarksresponsible AI
0
0 comments X

The pith

PaLM 2 raises quality on English, multilingual, and reasoning tasks while cutting inference time and compute compared to PaLM.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The report introduces PaLM 2 as a Transformer model trained with a mixture of objectives that outperforms its predecessor across language understanding, multilingual work, and reasoning benchmarks. It achieves these gains at multiple model sizes while also running faster during inference. A reader would care because the efficiency gains could allow wider use of capable models without proportional increases in hardware or energy costs. The work further shows that performance on responsible-AI checks remains stable and that toxicity can be adjusted at inference time without hurting other abilities. These results point to a practical advance in scaling language models.

Core claim

PaLM 2 is a new family of language models that, across sizes, produces measurably higher accuracy on downstream English and multilingual tasks and on reasoning suites such as BIG-Bench, while requiring less compute per token at inference time than the original PaLM.

What carries the argument

Mixture-of-objectives training on a Transformer backbone that jointly optimizes for language modeling, translation, and reasoning signals.

If this is right

  • Large gains on BIG-Bench and other reasoning benchmarks hold across model sizes.
  • Faster inference enables more natural, lower-latency user interactions.
  • Lower compute per token supports broader deployment of the models.
  • Performance on responsible-AI evaluations stays stable while allowing inference-time toxicity control.
  • The same efficiency pattern appears in both pre-trained and fine-tuned variants.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The efficiency pattern could lower the energy cost of running large models at scale.
  • Similar training mixtures might be tested on non-Transformer architectures to check whether the gains are architecture-specific.
  • If the multilingual improvements generalize, they could reduce the need for separate language-specific models.

Load-bearing premise

The chosen English, multilingual, and reasoning benchmarks plus the responsible-AI tests fully represent real-world use without undisclosed data filtering or post-training adjustments.

What would settle it

Running PaLM 2 and PaLM on a fresh set of tasks and hardware never seen during their development and finding no consistent quality or speed advantage for PaLM 2.

read the original abstract

We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is a Transformer-based model trained using a mixture of objectives. Through extensive evaluations on English and multilingual language, and reasoning tasks, we demonstrate that PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM. This improved efficiency enables broader deployment while also allowing the model to respond faster, for a more natural pace of interaction. PaLM 2 demonstrates robust reasoning capabilities exemplified by large improvements over PaLM on BIG-Bench and other reasoning tasks. PaLM 2 exhibits stable performance on a suite of responsible AI evaluations, and enables inference-time control over toxicity without additional overhead or impact on other capabilities. Overall, PaLM 2 achieves state-of-the-art performance across a diverse set of tasks and capabilities. When discussing the PaLM 2 family, it is important to distinguish between pre-trained models (of various sizes), fine-tuned variants of these models, and the user-facing products that use these models. In particular, user-facing products typically include additional pre- and post-processing steps. Additionally, the underlying models may evolve over time. Therefore, one should not expect the performance of user-facing products to exactly match the results reported in this report.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces PaLM 2, a Transformer-based language model trained using a mixture of objectives. It claims superior multilingual and reasoning capabilities, greater compute efficiency, and faster inference relative to PaLM, supported by extensive evaluations showing significantly improved quality on English, multilingual, and reasoning benchmarks (including large gains on BIG-Bench) across model sizes, plus stable performance on responsible-AI evaluations and inference-time toxicity control.

Significance. If the performance gains are genuine and stem from the mixture-of-objectives training rather than data overlap or undisclosed adjustments, the work advances understanding of efficient scaling for large language models and demonstrates practical benefits for deployment. The broad evaluation suite covering reasoning, multilingual, and responsible-AI tasks is a strength, though the high-level reporting limits replicability.

major comments (2)
  1. [Evaluations and Training sections] The manuscript provides no description of training data sources, decontamination procedures, or explicit confirmation that benchmark test sets (e.g., BIG-Bench) were excluded from the pretraining mixture. This is load-bearing for the central claim of 'significantly improved quality on downstream tasks' and 'large improvements over PaLM on BIG-Bench' because gains could arise from data contamination rather than the new training approach.
  2. [Abstract and Efficiency discussion] Quantitative details on inference efficiency (e.g., latency, throughput, or FLOPs comparisons to PaLM) and the specific mixture weights or model-size variants are absent from the high-level descriptions. These omissions undermine evaluation of the 'faster and more efficient inference' and 'more compute-efficient' claims, which are central to the contribution.
minor comments (1)
  1. [Abstract] The distinction between pre-trained models, fine-tuned variants, and user-facing products is noted but could be clarified with explicit mapping of which reported results apply to base models versus products.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for their detailed review and valuable suggestions. We address the major comments below and have updated the manuscript accordingly where feasible.

read point-by-point responses

  1. Referee: [Evaluations and Training sections] The manuscript provides no description of training data sources, decontamination procedures, or explicit confirmation that benchmark test sets (e.g., BIG-Bench) were excluded from the pretraining mixture. This is load-bearing for the central claim of 'significantly improved quality on downstream tasks' and 'large improvements over PaLM on BIG-Bench' because gains could arise from data contamination rather than the new training approach.

    Authors: We appreciate this important point. Due to the proprietary nature of the training data, we are unable to provide a full description of the data sources. However, we confirm that the pretraining mixture was carefully curated to exclude evaluation benchmarks, including those in BIG-Bench, using standard decontamination techniques. We have added a clarification in the Training section of the revised manuscript to explicitly state that benchmark test sets were not included in pretraining. This addresses the concern regarding potential data contamination. revision: partial

  2. Referee: [Abstract and Efficiency discussion] Quantitative details on inference efficiency (e.g., latency, throughput, or FLOPs comparisons to PaLM) and the specific mixture weights or model-size variants are absent from the high-level descriptions. These omissions undermine evaluation of the 'faster and more efficient inference' and 'more compute-efficient' claims, which are central to the contribution.

    Authors: We agree that providing more quantitative details would strengthen the manuscript. In the revised version, we have included specific comparisons of inference latency and throughput for PaLM 2 versus PaLM, along with details on the mixture-of-objectives weights and the different model size variants used in our experiments. These additions are now present in the Efficiency discussion section. revision: yes

standing simulated objections not resolved
  • Full disclosure of training data sources and exact compositions, which remain proprietary.

Circularity Check

0 steps flagged

No circularity: empirical results on external benchmarks

full rationale

The PaLM 2 technical report presents training details and measured performance on public external benchmarks (BIG-Bench, English/multilingual/reasoning suites). No load-bearing step reduces a claimed prediction or first-principles result to a quantity defined by the authors' own fitted parameters, self-citations, or ansatz. Distinctions between pre-trained models, fine-tuned variants, and user-facing products are explicit and do not create self-definition. Central claims rest on independent evaluation outcomes rather than internal re-labeling of inputs.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

This is an empirical engineering report rather than a derivation; the central claims rest on undisclosed choices of training data mixture, model scale, and evaluation protocols that function as free parameters. No new physical or mathematical axioms are introduced.

free parameters (2)
  • training objective mixture weights
    The mixture of objectives is stated but the relative weights and exact objectives are not quantified in the provided abstract.
  • model size variants
    Multiple sizes are evaluated but exact parameter counts and training compute budgets are not specified here.
axioms (1)
  • domain assumption Standard scaling assumptions in large language model training hold for the new mixture of objectives.
    The report assumes that prior scaling laws and Transformer training practices transfer without major modification.

pith-pipeline@v0.9.0 · 6073 in / 1344 out tokens · 55080 ms · 2026-05-12T11:54:22.599574+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 41 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents

    cs.LG 2026-05 unverdicted novelty 7.0

    HyperEyes uses a dual-grained RL framework with parallel tool actions and efficiency rewards to achieve 9.9% higher accuracy and 5.3x fewer tool calls than prior open-source multimodal agents.

  2. Logic-Regularized Verifier Elicits Reasoning from LLMs

    cs.CL 2026-05 unverdicted novelty 7.0

    LOVER creates an unsupervised logic-regularized verifier that reaches 95% of supervised verifier performance on reasoning tasks across 10 datasets.

  3. Adaptive Selection of LoRA Components in Privacy-Preserving Federated Learning

    cs.LG 2026-05 unverdicted novelty 7.0

    AS-LoRA adaptively chooses which LoRA factor to update per layer and round using a curvature-aware second-order score, eliminating reconstruction error floors and improving performance in DP federated learning.

  4. E-MIA: Exam-Style Black-Box Membership Inference Attacks against RAG Systems

    cs.CR 2026-05 unverdicted novelty 7.0

    E-MIA converts document details into four types of exam questions and aggregates the RAG's answers into a membership score that separates member and non-member documents better than prior similarity-based or probe-bas...

  5. InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees

    cs.LG 2026-05 unverdicted novelty 7.0

    InvEvolve evolves white-box inventory policies from LLMs with statistical safety guarantees and outperforms classical and deep learning methods on synthetic and real retail data.

  6. To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning

    cs.AI 2026-04 conditional novelty 7.0

    Unembedding collapse in transformers prevents distinguishing unseen tokens in symbolic reasoning, but targeted interventions restore generalization.

  7. RoLegalGEC: Legal Domain Grammatical Error Detection and Correction Dataset for Romanian

    cs.CL 2026-04 unverdicted novelty 7.0

    RoLegalGEC is the first Romanian legal-domain dataset for grammatical error detection and correction, consisting of 350,000 examples, with evaluations of several neural models.

  8. Drift-AR: Single-Step Visual Autoregressive Generation via Anti-Symmetric Drifting

    cs.CV 2026-03 unverdicted novelty 7.0

    Drift-AR achieves 3.8-5.5x speedup in AR-diffusion image models by using entropy to enable entropy-informed speculative decoding and single-step (1-NFE) anti-symmetric drifting decoding.

  9. Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation

    cs.CV 2024-06 conditional novelty 7.0

    Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.

  10. Open X-Embodiment: Robotic Learning Datasets and RT-X Models

    cs.RO 2023-10 unverdicted novelty 7.0

    A collaborative dataset spanning 22 robots and 527 skills enables RT-X models that transfer capabilities across different robot embodiments.

  11. Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation

    cs.CV 2023-10 unverdicted novelty 7.0

    A new shared video-image tokenizer enables large language models to surpass diffusion models on standard visual generation benchmarks.

  12. Large Language Models as Optimizers

    cs.LG 2023-09 unverdicted novelty 7.0

    Large language models can optimize by being prompted with histories of past solutions and scores to propose better ones, producing prompts that raise accuracy up to 8% on GSM8K and 50% on Big-Bench Hard over human-des...

  13. LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

    cs.LG 2026-05 unverdicted novelty 6.0

    LoKA enables practical FP8 use in numerically sensitive large recommendation models via online profiling of activations, reusable model modifications for stability, and dynamic kernel dispatching.

  14. LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

    cs.LG 2026-05 unverdicted novelty 6.0

    LoKA enables practical FP8 use in numerically sensitive large recommendation models via profiling, model adaptations, and runtime kernel orchestration.

  15. XPERT: Expert Knowledge Transfer for Effective Training of Language Models

    cs.CL 2026-05 unverdicted novelty 6.0

    XPERT extracts and reuses cross-domain expert knowledge from pre-trained MoE LLMs via inference analysis and tensor decomposition to improve performance and convergence in downstream language model training.

  16. HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents

    cs.LG 2026-05 unverdicted novelty 6.0

    HyperEyes presents a parallel multimodal search agent using dual-grained efficiency-aware RL with a new TRACE reward and IMEB benchmark, claiming 9.9% higher accuracy and 5.3x fewer tool calls than prior open-source agents.

  17. Towards Reliable LLM Evaluation: Correcting the Winner's Curse in Adaptive Benchmarking

    stat.ML 2026-05 unverdicted novelty 6.0

    SIREN corrects winner's curse bias in adaptive LLM benchmarking via selection-aware repeated splits and bootstrap for valid procedure-level confidence intervals.

  18. InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees

    cs.LG 2026-05 unverdicted novelty 6.0

    InvEvolve uses LLMs and RL to generate certified inventory policies that outperform classical and deep learning methods on synthetic and real data while providing multi-period performance guarantees.

  19. Breaking Lock-In: Preserving Steerability under Low-Data VLA Post-Training

    cs.RO 2026-04 unverdicted novelty 6.0

    DeLock mitigates lock-in in low-data VLA post-training via visual grounding preservation and test-time contrastive prompt guidance, outperforming baselines across eight evaluations while matching data-heavy generalist...

  20. EMMA: End-to-End Multimodal Model for Autonomous Driving

    cs.CV 2024-10 unverdicted novelty 6.0

    EMMA is an end-to-end multimodal LLM that converts camera data into trajectories, objects, and road graphs via text prompts and reports state-of-the-art motion planning on nuScenes plus competitive detection results on Waymo.

  21. LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code

    cs.SE 2024-03 unverdicted novelty 6.0

    LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.

  22. Corrective Retrieval Augmented Generation

    cs.CL 2024-01 unverdicted novelty 6.0

    CRAG improves RAG robustness via a retrieval quality evaluator that triggers web augmentation and a decompose-recompose filter to focus on relevant information, yielding better results on short- and long-form generati...

  23. Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations

    cs.AI 2023-12 conditional novelty 6.0

    Math-Shepherd is an automatically trained process reward model that scores solution steps to verify and reinforce LLMs, lifting Mistral-7B from 77.9% to 89.1% on GSM8K and 28.6% to 43.5% on MATH.

  24. Video-LLaVA: Learning United Visual Representation by Alignment Before Projection

    cs.CV 2023-11 unverdicted novelty 6.0

    Video-LLaVA creates a unified visual representation for images and videos via pre-projection alignment, enabling mutual enhancement from joint training and strong results on image and video benchmarks.

  25. Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models

    eess.AS 2023-11 unverdicted novelty 6.0

    Qwen-Audio trains a unified model on diverse audio and tasks with hierarchical tags to enable strong zero-shot performance on audio understanding benchmarks and multi-turn audio chat.

  26. Large Language Models Cannot Self-Correct Reasoning Yet

    cs.CL 2023-10 unverdicted novelty 6.0

    LLMs cannot reliably self-correct reasoning mistakes using only their internal capabilities and often degrade in performance without external feedback.

  27. MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models

    cs.CL 2023-09 conditional novelty 6.0

    Bootstrapping math questions via rewriting creates MetaMathQA; fine-tuning LLaMA-2 on it yields 66.4% on GSM8K for 7B and 82.3% for 70B, beating prior same-size models by large margins.

  28. GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts

    cs.AI 2023-09 unverdicted novelty 6.0

    GPTFuzz is a black-box fuzzing framework that mutates seed jailbreak templates to automatically generate effective attacks, achieving over 90% success rates on models including ChatGPT and Llama-2.

  29. Scaling Relationship on Learning Mathematical Reasoning with Large Language Models

    cs.CL 2023-08 unverdicted novelty 6.0

    Pre-training loss predicts LLM math reasoning better than parameter count; rejection sampling fine-tuning with diverse paths raises LLaMA-7B accuracy on GSM8K from 35.9% with SFT to 49.3%.

  30. Textbooks Are All You Need

    cs.CL 2023-06 unverdicted novelty 6.0

    A 1.3B-parameter code model trained on 7B tokens of curated textbook and synthetic data achieves 50.6% on HumanEval, indicating data quality can enable strong performance at small scale.

  31. MiniLLM: On-Policy Distillation of Large Language Models

    cs.CL 2023-06 conditional novelty 6.0

    MiniLLM distills large language models into smaller ones via reverse KL divergence and on-policy optimization, yielding higher-quality responses with lower exposure bias than standard KD baselines.

  32. Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

    cs.CL 2023-06 accept novelty 6.0

    GPT-4 as an LLM judge achieves over 80% agreement with human preferences on MT-Bench and Chatbot Arena, matching human agreement levels and providing a scalable evaluation method.

  33. Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models

    cs.AI 2025-03 unverdicted novelty 5.0

    The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.

  34. Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations

    cs.CL 2023-12 unverdicted novelty 5.0

    Llama Guard is an instruction-tuned Llama2-7b model that performs multi-class safety classification on prompts and responses, matching or exceeding existing moderation tools on benchmarks while supporting taxonomy cus...

  35. VectraYX-Nano: A 42M-Parameter Spanish Cybersecurity Language Model with Curriculum Learning and Native Tool Use

    cs.CL 2026-05 unverdicted novelty 4.0

    VectraYX-Nano is a 42M-parameter Spanish cybersecurity LLM trained with curriculum learning and native MCP tool use, achieving 0.78 conversational gate and improved tool selection with denser data.

  36. UnAC: Adaptive Visual Prompting with Abstraction and Stepwise Checking for Complex Multimodal Reasoning

    cs.CV 2026-05 unverdicted novelty 4.0

    UnAC improves LMM performance on visual reasoning benchmarks by combining adaptive visual prompting, image abstraction, and gradual self-checking.

  37. MedThink: Enhancing Diagnostic Accuracy in Small Models via Teacher-Guided Reasoning Correction

    cs.CY 2026-04 unverdicted novelty 4.0

    MedThink, a two-stage teacher-guided reasoning correction distillation framework, boosts small language models' medical diagnostic accuracy by up to 12.7% on benchmarks and achieves 56.4% on a gastroenterology dataset.

  38. Gemma: Open Models Based on Gemini Research and Technology

    cs.CL 2024-03 accept novelty 4.0

    Gemma introduces open 2B and 7B LLMs derived from Gemini technology that beat comparable open models on 11 of 18 text tasks and come with safety assessments.

  39. Gemma 2: Improving Open Language Models at a Practical Size

    cs.CL 2024-07 conditional novelty 3.0

    Gemma 2 models achieve leading performance at their sizes by combining established Transformer modifications with knowledge distillation for the 2B and 9B variants.

  40. Large Language Models: A Survey

    cs.CL 2024-02 accept novelty 3.0

    The paper surveys key large language models, their training methods, datasets, evaluation benchmarks, and future research directions in the field.

  41. A Survey of Large Language Models

    cs.CL 2023-03 accept novelty 3.0

    This survey reviews the background, key techniques, and evaluation methods for large language models, emphasizing emergent abilities that appear at large scales.

Reference graph

Works this paper leans on

286 extracted references · 286 canonical work pages · cited by 38 Pith papers · 28 internal anchors

  1. [1]

    Persistent anti-muslim bias in large language models

    Abid, A., Farooqi, M., and Zou, J. Persistent anti-muslim bias in large language models. arXiv preprint arXiv:2101.05783, 2021. URL https://arxiv.org/abs/2101.05783

    work page arXiv 2021

  2. [2]

    Akhbardeh, F., Arkhangorodsky, A., Biesialska, M., Bojar, O., Chatterjee, R., Chaudhary, V., Costa-jussa, M. R., Espa \ n a-Bonet, C., Fan, A., Federmann, C., Freitag, M., Graham, Y., Grundkiewicz, R., Haddow, B., Harter, L., Heafield, K., Homan, C., Huck, M., Amponsah-Kaakyire, K., Kasai, J., Khashabi, D., Knight, K., Kocmi, T., Koehn, P., Lourie, N., Mo...

    work page 2021

  3. [3]

    Guide to fair pay, 2023

    Appen. Guide to fair pay, 2023. URL https://success.appen.com/hc/en-us/articles/9557008940941-Guide-to-Fair-Pay

    work page arXiv 2023

  4. [5]

    Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

    Bai, Y., Jones, A., Ndousse, K., Askell, A., Chen, A., DasSarma, N., Drain, D., Fort, S., Ganguli, D., Henighan, T., Joseph, N., Kadavath, S., Kernion, J., Conerly, T., El-Showk, S., Elhage, N., Hatfield-Dodds, Z., Hernandez, D., Hume, T., Johnston, S., Kravec, S., Lovitt, L., Nanda, N., Olsson, C., Amodei, D., Brown, T., Clark, J., McCandlish, S., Olah, ...

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  5. [6]

    X., Macherey, K., Krikun, M., Wang, P., Gutkin, A., Shah, A., Huang, Y., Chen, Z., Wu, Y., and Hughes, M

    Bapna, A., Caswell, I., Kreutzer, J., Firat, O., van Esch, D., Siddhant, A., Niu, M., Baljekar, P., Garcia, X., Macherey, W., Breiner, T., Axelrod, V., Riesa, J., Cao, Y., Chen, M. X., Macherey, K., Krikun, M., Wang, P., Gutkin, A., Shah, A., Huang, Y., Chen, Z., Wu, Y., and Hughes, M. Building machine translation systems for the next thousand languages. ...

    work page arXiv 2022

  6. [7]

    Pathways: Asynchronous distributed dataflow for ml

    Barham, P., Chowdhery, A., Dean, J., Ghemawat, S., Hand, S., Hurt, D., Isard, M., Lim, H., Pang, R., Roy, S., et al. Pathways: Asynchronous distributed dataflow for ml. Proceedings of Machine Learning and Systems, 4: 0 430--449, 2022

    work page 2022

  7. [8]

    Fairness and machine learning limitations and opportunities

    Barocas, S., Hardt, M., and Narayanan, A. Fairness and machine learning limitations and opportunities. 2017

    work page 2017

  8. [9]

    In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society

    Barocas, S., Guo, A., Kamar, E., Krones, J., Morris, M. R., Vaughan, J. W., Wadsworth, W. D., and Wallach, H. Designing disaggregated evaluations of ai systems: Choices, considerations, and tradeoffs. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, AIES '21, pp.\ 368–378, New York, NY, USA, 2021. Association for Computing Machin...

    work page doi:10.1145/3461702.3462610 2021

  9. [10]

    Bender, E. M. and Friedman, B. Data statements for natural language processing: Toward mitigating system bias and enabling better science. Transactions of the Association for Computational Linguistics, 6: 0 587--604, 2018. doi:10.1162/tacl_a_00041. URL https://aclanthology.org/Q18-1041

    work page doi:10.1162/tacl_a_00041 2018

  10. [11]

    Semantic parsing on F reebase from question-answer pairs

    Berant, J., Chou, A., Frostig, R., and Liang, P. Semantic parsing on F reebase from question-answer pairs. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pp.\ 1533--1544, Seattle, Washington, USA, October 2013. Association for Computational Linguistics. URL https://aclanthology.org/D13-1160

    work page 2013

  11. [12]

    Re-contextualizing fairness in NLP : The case of india

    Bhatt, S., Dev, S., Talukdar, P., Dave, S., and Prabhakaran, V. Re-contextualizing fairness in NLP : The case of india. September 2022. URL https://arxiv.org/abs/2209.12226

    work page arXiv 2022

  12. [13]

    Piqa: Reasoning about physical commonsense in natural language

    Bisk, Y., Zellers, R., Gao, J., Choi, Y., et al. Piqa: Reasoning about physical commonsense in natural language. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.\ 7432--7439, 2020

    work page 2020

  13. [14]

    L., Barocas, S., Daum \'e , III, H., and Wallach, H

    Blodgett, S. L., Barocas, S., Daum \'e , III, H., and Wallach, H. Language (technology) is power: A critical survey of ``bias'' in NLP . May 2020. URL https://arxiv.org/abs/2005.14050

    work page arXiv 2020

  14. [15]

    L., Lopez, G., Olteanu, A., Sim, R., and Wallach, H

    Blodgett, S. L., Lopez, G., Olteanu, A., Sim, R., and Wallach, H. Stereotyping N orwegian salmon: An inventory of pitfalls in fairness benchmark datasets. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pp.\ 1004--...

    work page doi:10.18653/v1/2021.acl-long.81 2021

  15. [16]

    Nuanced metrics for measuring unintended bias with real data for text classification

    Borkan, D., Dixon, L., Sorensen, J., Thain, N., and Vasserman, L. Nuanced metrics for measuring unintended bias with real data for text classification, 2019. URL https://arxiv.org/abs/1903.04561

    work page arXiv 2019

  16. [17]

    Bowman, S. R. and Dahl, G. E. What will it take to fix benchmarking in natural language understanding?, 2021

    work page 2021

  17. [18]

    J., Leary, C., Maclaurin, D., Necula, G., Paszke, A., Vander P las, J., Wanderman- M ilne, S., and Zhang, Q

    Bradbury, J., Frostig, R., Hawkins, P., Johnson, M. J., Leary, C., Maclaurin, D., Necula, G., Paszke, A., Vander P las, J., Wanderman- M ilne, S., and Zhang, Q. JAX : composable transformations of P ython+ N um P y programs, 2018. URL http://github.com/google/jax

    work page 2018

  18. [19]

    Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., Hesse, C., Chen, M., Sigler, E., Litwin, M., Gray, S., Chess, B., Clark, J., Berner, C., McCandlish, S., Radford, A.,...

    work page 1901

  19. [20]

    The secret sharer: Evaluating and testing unintended memorization in neural networks

    Carlini, N., Liu, C., Erlingsson, \'U ., Kos, J., and Song, D. The secret sharer: Evaluating and testing unintended memorization in neural networks. In USENIX Security Symposium, volume 267, 2019

    work page 2019

  20. [21]

    B., Song, D., Erlingsson, U., et al

    Carlini, N., Tramer, F., Wallace, E., Jagielski, M., Herbert-Voss, A., Lee, K., Roberts, A., Brown, T. B., Song, D., Erlingsson, U., et al. Extracting training data from large language models. In USENIX Security Symposium, volume 6, 2021

    work page 2021

  21. [23]

    J., Hale, P., and Wachs, F

    Casad, B. J., Hale, P., and Wachs, F. L. Stereotype threat among girls: Differences by gender identity and math education context, 2017

    work page 2017

  22. [24]

    Question directed graph attention network for numerical reasoning over text

    Chen, K., Xu, W., Cheng, X., Xiaochuan, Z., Zhang, Y., Song, L., Wang, T., Qi, Y., and Chu, W. Question directed graph attention network for numerical reasoning over text. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp.\ 6759--6768, Online, November 2020. Association for Computational Linguistics. doi...

    work page doi:10.18653/v1/2020.emnlp-main.549 2020

  23. [26]

    PaLM: Scaling Language Modeling with Pathways

    Chowdhery, A., Narang, S., Devlin, J., Bosma, M., Mishra, G., Chung, H. W., Sutton, C., Gehrmann, S., Schuh, P., et al. Pa LM : S caling language modeling with P athways. arXiv preprint arXiv:2204.02311, 2022. URL https://arxiv.org/abs/2204.02311

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  24. [27]

    Scaling Instruction-Finetuned Language Models

    Chung, H. W., Hou, L., Longpre, S., Zoph, B., Tay, Y., Fedus, W., Li, Y., Wang, X., Dehghani, M., Brahma, S., Webson, A., Gu, S. S., Dai, Z., Suzgun, M., Chen, X., Chowdhery, A., Castro-Ros, A., Pellat, M., Robinson, K., Valter, D., Narang, S., Mishra, G., Yu, A., Zhao, V., Huang, Y., Dai, A., Yu, H., Petrov, S., Chi, E. H., Dean, J., Devlin, J., Roberts,...

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  25. [28]

    H., Choi, E., Collins, M., Garrette, D., Kwiatkowski, T., Nikolaev, V., and Palomaki, J

    Clark, J. H., Choi, E., Collins, M., Garrette, D., Kwiatkowski, T., Nikolaev, V., and Palomaki, J. T y D i QA : A benchmark for information-seeking question answering in typologically diverse languages. TACL, 2020. URL https://aclanthology.org/2020.tacl-1.30

    work page 2020

  26. [29]

    Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge

    Clark, P., Cowhey, I., Etzioni, O., Khot, T., Sabharwal, A., Schoenick, C., and Tafjord, O. Think you have solved question answering? T ry arc, the AI2 reasoning challenge. arXiv preprint arXiv:1803.05457, 2018. URL https://arxiv.org/abs/1803.05457

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  27. [31]

    Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics, 1989

    Crenshaw, K. Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics, 1989

    work page 1989

  28. [32]

    Dai, A. M. and Le, Q. V. Semi-supervised sequence learning. In Cortes, C., Lawrence, N., Lee, D., Sugiyama, M., and Garnett, R. (eds.), Advances in Neural Information Processing Systems, volume 28. Curran Associates, Inc., 2015. URL https://proceedings.neurips.cc/paper_files/paper/2015/file/7137debd45ae4d0ab9aa953017286b20-Paper.pdf

    work page 2015

  29. [33]

    Daniels, P. T. and Bright, W. The world's writing systems. Oxford University Press on Demand, 1996

    work page 1996

  30. [34]

    Denton, E., Hanna, A., Amironesei, R., Smart, A., Nicole, H., and Scheuerman, M. K. Bringing the people back in: Contesting benchmark machine learning datasets, 2020

    work page 2020

  31. [35]

    M., and Chang, K.-W

    Dev, S., Monajatipoor, M., Ovalle, A., Subramonian, A., Phillips, J. M., and Chang, K.-W. Harms of gender exclusivity and challenges in non-binary representation in language technologies, 2021 a . URL https://arxiv.org/abs/2108.12084

    work page arXiv 2021

  32. [36]

    On measures of biases and harms in NLP

    Dev, S., Sheng, E., Zhao, J., Amstutz, A., Sun, J., Hou, Y., Sanseverino, M., Kim, J., Nishi, A., Peng, N., and Chang, K.-W. On measures of biases and harms in NLP . August 2021 b . URL https://arxiv.org/abs/2108.03362

    work page arXiv 2021

  33. [37]

    BERT : P re-training of deep bidirectional transformers for language understanding

    Devlin, J., Chang, M.-W., Lee, K., and Toutanova, K. BERT : P re-training of deep bidirectional transformers for language understanding. NAACL, 2019. URL https://aclanthology.org/N19-1423

    work page 2019

  34. [38]

    D., Rosen, R., Baker, D

    Diaz, M., Kivlichan, I. D., Rosen, R., Baker, D. K., Amironesei, R., Prabhakaran, V., and Denton, E. CrowdWorkSheets : Accounting for individual and collective identities underlying crowdsourced dataset annotation. June 2022. URL https://arxiv.org/abs/2206.08931

    work page arXiv 2022

  35. [39]

    Build it break it fix it for dialogue safety: Robustness from adversarial human attack

    Dinan, E., Humeau, S., Chintagunta, B., and Weston, J. Build it break it fix it for dialogue safety: Robustness from adversarial human attack. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pp.\ 4537--4546, Hong Kong, China,...

    work page doi:10.18653/v1/d19-1461 2019

  36. [40]

    Documenting large webtext corpora: A case study on the colossal clean crawled corpus, 2021

    Dodge, J., Sap, M., Marasović, A., Agnew, W., Ilharco, G., Groeneveld, D., Mitchell, M., and Gardner, M. Documenting large webtext corpora: A case study on the colossal clean crawled corpus, 2021

    work page 2021

  37. [41]

    arXiv preprint arXiv:2112.06905 , year =

    Du , N., Huang , Y., Dai , A. M., Tong , S., Lepikhin , D., Xu , Y., Krikun , M., Zhou , Y., Yu , A. W., Firat , O., Zoph , B., Fedus , L., Bosma , M., Zhou , Z., Wang , T., Wang , Y. E., Webster , K., Pellat , M., Robinson , K., Meier-Hellstern , K., Duke , T., Dixon , L., Zhang , K., Le , Q. V., Wu , Y., Chen , Z., and Cui , C. GLaM: Efficient Scaling o...

    work page arXiv 2022

  38. [42]

    DROP : A reading comprehension benchmark requiring discrete reasoning over paragraphs

    Dua, D., Wang, Y., Dasigi, P., Stanovsky, G., Singh, S., and Gardner, M. DROP : A reading comprehension benchmark requiring discrete reasoning over paragraphs. In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) , pp.\ 2368--237...

    work page doi:10.18653/v1/n19-1246 2019

  39. [44]

    Experts, errors, and context: A large-scale study of human evaluation for machine translation

    Freitag, M., Foster, G., Grangier, D., Ratnakar, V., Tan, Q., and Macherey, W. Experts, errors, and context: A large-scale study of human evaluation for machine translation. Transactions of the Association for Computational Linguistics, 9: 0 1460--1474, 2021. doi:10.1162/tacl_a_00437. URL https://aclanthology.org/2021.tacl-1.87

    work page doi:10.1162/tacl_a_00437 2021

  40. [45]

    Freitag, M., Rei, R., Mathur, N., Lo, C.-k., Stewart, C., Avramidis, E., Kocmi, T., Foster, G., Lavie, A., and Martins, A. F. T. Results of WMT 22 metrics shared task: Stop using BLEU -- neural metrics are better and more robust. In Proceedings of the Seventh Conference on Machine Translation (WMT), pp.\ 46--68, Abu Dhabi, United Arab Emirates (Hybrid), D...

    work page 2022

  41. [46]

    Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned

    Ganguli, D., Lovitt, L., Kernion, J., Askell, A., Bai, Y., Kadavath, S., Mann, B., Perez, E., Schiefer, N., Ndousse, K., Jones, A., Bowman, S., Chen, A., Conerly, T., DasSarma, N., Drain, D., Elhage, N., El-Showk, S., Fort, S., Hatfield-Dodds, Z., Henighan, T., Hernandez, D., Hume, T., Jacobson, J., Johnston, S., Kravec, S., Olsson, C., Ringer, S., Tran-J...

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  42. [47]

    Word embeddings quantify 100 years of gender and ethnic stereotypes

    Garg, N., Schiebinger, L., Jurafsky, D., and Zou, J. Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences, 115 0 (16): 0 E3635--E3644, 2018. doi:10.1073/pnas.1720347115. URL https://www.pnas.org/doi/abs/10.1073/pnas.1720347115

    work page doi:10.1073/pnas.1720347115 2018

  43. [48]

    Handling bias in toxic speech detection: A survey

    Garg, T., Masud, S., Suresh, T., and Chakraborty, T. Handling bias in toxic speech detection: A survey. January 2022. URL https://arxiv.org/abs/2202.00126

    work page arXiv 2022

  44. [49]

    W., Wallach, H., au2, H

    Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., au2, H. D. I., and Crawford, K. Datasheets for datasets, 2021

    work page 2021

  45. [50]

    Gehman, S., Gururangan, S., Sap, M., Choi, Y., and Smith, N. A. R eal T oxicity P rompts: Evaluating neural toxic degeneration in language models. In Findings of the Association for Computational Linguistics: EMNLP 2020, pp.\ 3356--3369, Online, November 2020. Association for Computational Linguistics. doi:10.18653/v1/2020.findings-emnlp.301. URL https://...

    work page doi:10.18653/v1/2020.findings-emnlp.301 2020

  46. [52]

    Your AI pair programmer, October 2021

    Github. Your AI pair programmer, October 2021

    work page 2021

  47. [53]

    Improving alignment of dialogue agents via targeted human judgements

    Glaese, A., McAleese, N., Trębacz, M., Aslanides, J., Firoiu, V., Ewalds, T., Rauh, M., Weidinger, L., Chadwick, M., Thacker, P., Campbell-Gillingham, L., Uesato, J., Huang, P.-S., Comanescu, R., Yang, F., See, A., Dathathri, S., Greig, R., Chen, C., Fritz, D., Elias, J. S., Green, R., Mokrá, S., Fernando, N., Wu, B., Foley, R., Young, S., Gabriel, I., Is...

    work page internal anchor Pith review arXiv 2022

  48. [54]

    Intrinsic bias metrics do not correlate with application bias

    Goldfarb-Tarrant, S., Marchant, R., Mu \ n oz S \'a nchez, R., Pandya, M., and Lopez, A. Intrinsic bias metrics do not correlate with application bias. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pp.\ 1926--194...

    work page doi:10.18653/v1/2021.acl-long.150 1926

  49. [55]

    Our principles, 2018

    Google. Our principles, 2018. URL https://ai.google/responsibility/principles/. Accessed May 16, 2023

    work page 2018

  50. [56]

    Generative ai prohibited use policy, 2023 a

    Google. Generative ai prohibited use policy, 2023 a . URL https://policies.google.com/terms/generative-ai/use-policy. Accessed May 16, 2023

    work page 2023

  51. [57]

    Palm api and makersuite additional terms of service, 2023 b

    Google. Palm api and makersuite additional terms of service, 2023 b . URL https://developers.generativeai.google/terms. Accessed May 16, 2023

    work page 2023

  52. [58]

    Is your toxicity my toxicity? E xploring the impact of rater identity on toxicity annotation

    Goyal, N., Kivlichan, I., Rosen, R., and Vasserman, L. Is your toxicity my toxicity? E xploring the impact of rater identity on toxicity annotation. May 2022. URL https://arxiv.org/abs/2205.00501

    work page arXiv 2022

  53. [59]

    Generating sequences with recurrent neural networks, 2014

    Graves, A. Generating sequences with recurrent neural networks, 2014

    work page 2014

  54. [60]

    Towards a critical race methodology in algorithmic fairness

    Hanna, A., Denton, E., Smart, A., and Smith-Loud, J. Towards a critical race methodology in algorithmic fairness. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, FAT* '20, pp.\ 501–512, New York, NY, USA, 2020. Association for Computing Machinery. ISBN 9781450369367. doi:10.1145/3351095.3372826. URL https://doi.org/10....

    work page doi:10.1145/3351095.3372826 2020

  55. [61]

    S., Mubasshir, K., Li, Y.-F., Kang, Y.-B., Rahman, M

    Hasan, T., Bhattacharjee, A., Islam, M. S., Mubasshir, K., Li, Y.-F., Kang, Y.-B., Rahman, M. S., and Shahriyar, R. XL -sum: Large-scale multilingual abstractive summarization for 44 languages. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pp.\ 4693--4703, Online, August 2021. Association for Computational Linguistics. doi...

    work page doi:10.18653/v1/2021.findings-acl.413 2021

  56. [62]

    A., Burns, K., Saenko, K., Darrell, T., and Rohrbach, A

    Hendricks, L. A., Burns, K., Saenko, K., Darrell, T., and Rohrbach, A. Women also snowboard: Overcoming bias in captioning models (extended abstract), 2018

    work page 2018

  57. [64]

    Neural Computation 9, 1735–1780

    Hochreiter, S. and Schmidhuber, J. Long Short-Term Memory . Neural Computation, 9 0 (8): 0 1735--1780, 11 1997. ISSN 0899-7667. doi:10.1162/neco.1997.9.8.1735. URL https://doi.org/10.1162/neco.1997.9.8.1735

    work page doi:10.1162/neco.1997.9.8.1735 1997

  58. [65]

    Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., Casas, D. d. L., Hendricks, L. A., et al. Training compute-optimal large language models. NeurIPS, 2022. URL https://arxiv.org/abs/2203.15556

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  59. [66]

    Universal Language Model Fine-tuning for Text Classification

    Howard, J. and Ruder, S. Universal language model fine-tuning for text classification. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp.\ 328--339, Melbourne, Australia, July 2018. Association for Computational Linguistics. doi:10.18653/v1/P18-1031. URL https://aclanthology.org/P18-1031

    work page doi:10.18653/v1/p18-1031 2018

  60. [67]

    and Collins, E

    Hsiao, S. and Collins, E. Try bard and share your feedback. https://blog.google/technology/ai/try-bard/, March 2023. Accessed: 2023-5-5

    work page 2023

  61. [68]

    LoRA: Low-Rank Adaptation of Large Language Models

    Hu, E. J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., and Chen, W. LoRA : Low-Rank adaptation of large language models. June 2021. URL https://arxiv.org/abs/2106.09685

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  62. [70]

    Jacobs, A. Z. and Wallach, H. Measurement and fairness. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, FAccT '21, pp.\ 375–385, New York, NY, USA, 2021. Association for Computing Machinery. ISBN 9781450383097. doi:10.1145/3442188.3445901. URL https://doi.org/10.1145/3442188.3445901

    work page doi:10.1145/3442188.3445901 2021

  63. [72]

    ACM Comput

    Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., and Fung, P. Survey of hallucination in natural language generation. ACM Computing Surveys , 55 0 (12): 0 1--38, mar 2023. doi:10.1145/3571730. URL https://doi.org/10.1145

    work page doi:10.1145/3571730 2023

  64. [73]

    Toxic comment classification challenge, 2018

    Jigsaw. Toxic comment classification challenge, 2018. URL https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge

    work page 2018

  65. [74]

    Exploring the role of human raters in creating nlp datasets, 2019 a

    Jigsaw. Exploring the role of human raters in creating nlp datasets, 2019 a . URL https://medium.com/jigsaw/creating-labeled-datasets-and-exploring-the-role-of-human-raters-56367b6db298

    work page 2019

  66. [75]

    Jigsaw multilingual toxic comment classification, 2019 b

    Jigsaw. Jigsaw multilingual toxic comment classification, 2019 b . URL https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification

    work page 2019

  67. [76]

    T rivia QA : A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension

    Joshi, M., Choi, E., Weld, D., and Zettlemoyer, L. T rivia QA : A large scale distantly supervised challenge dataset for reading comprehension. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp.\ 1601--1611, Vancouver, Canada, July 2017. Association for Computational Linguistics. doi:10....

    work page doi:10.18653/v1/p17-1147 2017

  68. [77]

    P., Yoon, D

    Jouppi, N. P., Yoon, D. H., Kurian, G., Li, S., Patil, N., Laudon, J., Young, C., and Patterson, D. A domain-specific supercomputer for training deep neural networks. Communications of the ACM, 63 0 (7): 0 67--78, 2020

    work page 2020

  69. [79]

    The misgendering machines: Trans/hci implications of automatic gender recognition

    Keyes, O. The misgendering machines: Trans/hci implications of automatic gender recognition. Proc. ACM Hum.-Comput. Interact., 2 0 (CSCW), nov 2018. doi:10.1145/3274357. URL https://doi.org/10.1145/3274357

    work page doi:10.1145/3274357 2018

  70. [80]

    and Ney, H

    Kneser, R. and Ney, H. Improved backing-off for m-gram language modeling. In 1995 International Conference on Acoustics, Speech, and Signal Processing, volume 1, pp.\ 181--184 vol.1, 1995. doi:10.1109/ICASSP.1995.479394

    work page doi:10.1109/icassp.1995.479394 1995

  71. [81]

    Pretraining language models with human preferences

    Korbak, T., Shi, K., Chen, A., Bhalerao, R., Buckley, C. L., Phang, J., Bowman, S. R., and Perez, E. Pretraining language models with human preferences, 2023. URL https://arxiv.org/abs/2302.08582

    work page arXiv 2023

  72. [82]

    Quality at a glance: An audit of web-crawled multilingual datasets

    Kreutzer, J., Caswell, I., Wang, L., Wahab, A., van Esch, D., Ulzii-Orshikh, N., Tapo, A., Subramani, N., Sokolov, A., Sikasote, C., et al. Quality at a glance: An audit of web-crawled multilingual datasets. Transactions of the Association for Computational Linguistics, 10: 0 50--72, 2022

    work page 2022

  73. [83]

    https://aclanthology.org/ Q19-1026/

    Kwiatkowski, T., Palomaki, J., Redfield, O., Collins, M., Parikh, A., Alberti, C., Epstein, D., Polosukhin, I., Devlin, J., Lee, K., Toutanova, K., Jones, L., Kelcey, M., Chang, M.-W., Dai, A. M., Uszkoreit, J., Le, Q., and Petrov, S. Natural questions: A benchmark for question answering research. Transactions of the Association for Computational Linguist...

    work page doi:10.1162/tacl_a_00276 2019

  74. [84]

    Findings of the

    Ladhak, F., Durmus, E., Cardie, C., and McKeown, K. W iki L ingua: A new benchmark dataset for cross-lingual abstractive summarization. In Findings of the Association for Computational Linguistics: EMNLP 2020, pp.\ 4034--4048, Online, November 2020. Association for Computational Linguistics. doi:10.18653/v1/2020.findings-emnlp.360. URL https://aclantholog...

    work page doi:10.18653/v1/2020.findings-emnlp.360 2020

  75. [85]

    RACE : Large-scale R e A ding comprehension dataset from examinations

    Lai, G., Xie, Q., Liu, H., Yang, Y., and Hovy, E. RACE : Large-scale R e A ding comprehension dataset from examinations. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pp.\ 785--794, Copenhagen, Denmark, September 2017. Association for Computational Linguistics. doi:10.18653/v1/D17-1082. URL https://aclanthology...

    work page doi:10.18653/v1/d17-1082 2017

  76. [86]

    Welcome, singular "they"

    Lee, C. Welcome, singular "they". https://apastyle.apa.org/blog/singular-they, 2019. Accessed: 2022-11-18

    work page 2019

  77. [88]

    doi: 10.18653/v1/2021.emnlp-main.243

    Lester, B., Al-Rfou, R., and Constant, N. The power of scale for parameter-efficient prompt tuning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pp.\ 3045--3059, Online and Punta Cana, Dominican Republic, November 2021. Association for Computational Linguistics. doi:10.18653/v1/2021.emnlp-main.243. URL https:/...

    work page doi:10.18653/v1/2021.emnlp-main.243 2021

  78. [89]

    The winograd schema challenge

    Levesque, H., Davis, E., and Morgenstern, L. The winograd schema challenge. In Thirteenth international conference on the principles of knowledge representation and reasoning, 2012

    work page 2012

  79. [91]

    Holistic Evaluation of Language Models

    Liang, P., Bommasani, R., Lee, T., Tsipras, D., Soylu, D., Yasunaga, M., Zhang, Y., Narayanan, D., Wu, Y., Kumar, A., Newman, B., Yuan, B., Yan, B., Zhang, C., Cosgrove, C., Manning, C. D., R \'e , C., Acosta-Navas, D., Hudson, D. A., Zelikman, E., Durmus, E., Ladhak, F., Rong, F., Ren, H., Yao, H., Wang, J., Santhanam, K., Orr, L., Zheng, L., Yuksekgonul...

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  80. [92]

    W., Tay, Y ., Zhou, D., Le, Q

    Longpre, S., Hou, L., Vu, T., Webson, A., Chung, H. W., Tay, Y., Zhou, D., Le, Q. V., Zoph, B., Wei, J., and Roberts, A. The flan collection: Designing data and methods for effective instruction tuning. arXiv preprint arXiv:2301.13688, 2023

    work page arXiv 2023

Showing first 80 references.