[
  {
    "category_id": 1,
    "name": "Training and Scaling Experiments",
    "description": "Experiments focused on evaluating the impact of different training methods, parameter counts, and dataset sizes on model performance.",
    "children": [
      {
        "id": "360287970189639681",
        "label": "Paper",
        "properties": {
          "title": "Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone"
        }
      },
      {
        "id": "360287970189639682",
        "label": "Paper",
        "properties": {
          "title": "Averaging Weights Leads to Wider Optima and Better Generalization"
        }
      },
      {
        "id": "360287970189639687",
        "label": "Paper",
        "properties": {
          "title": "Language Models are Multilingual Chain-of-Thought Reasoners"
        }
      },
      {
        "id": "360287970189639718",
        "label": "Paper",
        "properties": {
          "title": "Unified Language Model Pre-training for Natural Language Understanding and Generation"
        }
      },
      {
        "id": "360287970189639720",
        "label": "Paper",
        "properties": {
          "title": "Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models"
        }
      },
      {
        "id": "360287970189639722",
        "label": "Paper",
        "properties": {
          "title": "MLS: A Large-Scale Multilingual Dataset for Speech Research"
        }
      },
      {
        "id": "360287970189639725",
        "label": "Paper",
        "properties": {
          "title": "InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning"
        }
      },
      {
        "id": "360287970189639727",
        "label": "Paper",
        "properties": {
          "title": "PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel"
        }
      },
      {
        "id": "360287970189639733",
        "label": "Paper",
        "properties": {
          "title": "Code Llama: Open Foundation Models for Code"
        }
      },
      {
        "id": "360287970189639761",
        "label": "Paper",
        "properties": {
          "title": "RoBERTa: A Robustly Optimized BERT Pretraining Approach"
        }
      },
      {
        "id": "360287970189639765",
        "label": "Paper",
        "properties": {
          "title": "Augmented Language Models: a Survey"
        }
      },
      {
        "id": "360287970189639766",
        "label": "Paper",
        "properties": {
          "title": "Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling"
        }
      },
      {
        "id": "360287970189639767",
        "label": "Paper",
        "properties": {
          "title": "Proximal Policy Optimization Algorithms"
        }
      },
      {
        "id": "360287970189639770",
        "label": "Paper",
        "properties": {
          "title": "GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints"
        }
      },
      {
        "id": "360287970189639774",
        "label": "Paper",
        "properties": {
          "title": "Reducing Activation Recomputation in Large Transformer Models"
        }
      },
      {
        "id": "360287970189639775",
        "label": "Paper",
        "properties": {
          "title": "SemDeDup: Data-efficient learning at web-scale through semantic deduplication"
        }
      },
      {
        "id": "360287970189639705",
        "label": "Paper",
        "properties": {
          "title": "GPQA: A Graduate-Level Google-Proof Q&A Benchmark"
        }
      },
      {
        "id": "360287970189639777",
        "label": "Paper",
        "properties": {
          "title": "RoFormer: Enhanced Transformer with Rotary Position Embedding"
        }
      },
      {
        "id": "360287970189639784",
        "label": "Paper",
        "properties": {
          "title": "ZeRO: Memory Optimizations Toward Training Trillion Parameter Models"
        }
      },
      {
        "id": "360287970189639785",
        "label": "Paper",
        "properties": {
          "title": "Branch-Train-Merge: Embarrassingly Parallel Training of Expert Language Models"
        }
      },
      {
        "id": "360287970189639796",
        "label": "Paper",
        "properties": {
          "title": "$\\infty$Bench: Extending Long Context Evaluation Beyond 100K Tokens"
        }
      },
      {
        "id": "360287970189639801",
        "label": "Paper",
        "properties": {
          "title": "The Falcon Series of Open Language Models"
        }
      },
      {
        "id": "360287970189639754",
        "label": "Paper",
        "properties": {
          "title": "The Stack: 3 TB of permissively licensed source code"
        }
      },
      {
        "id": "360287970189639804",
        "label": "Paper",
        "properties": {
          "title": "ZeRO-Offload: Democratizing Billion-Scale Model Training"
        }
      },
      {
        "id": "360287970189639808",
        "label": "Paper",
        "properties": {
          "title": "Mixture-of-Experts with Expert Choice Routing"
        }
      },
      {
        "id": "360287970189639815",
        "label": "Paper",
        "properties": {
          "title": "VideoChat: Chat-Centric Video Understanding"
        }
      },
      {
        "id": "360287970189639817",
        "label": "Paper",
        "properties": {
          "title": "Changing Answer Order Can Decrease MMLU Accuracy"
        }
      },
      {
        "id": "360287970189639819",
        "label": "Paper",
        "properties": {
          "title": "Scalable Extraction of Training Data from (Production) Language Models"
        }
      },
      {
        "id": "360287970189639825",
        "label": "Paper",
        "properties": {
          "title": "Don't Stop Pretraining: Adapt Language Models to Domains and Tasks"
        }
      },
      {
        "id": "360287970189639746",
        "label": "Paper",
        "properties": {
          "title": "Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time"
        }
      },
      {
        "id": "360287970189639827",
        "label": "Paper",
        "properties": {
          "title": "DocVQA: A Dataset for VQA on Document Images"
        }
      },
      {
        "id": "360287970189639828",
        "label": "Paper",
        "properties": {
          "title": "GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism"
        }
      },
      {
        "id": "360287970189639884",
        "label": "Paper",
        "properties": {
          "title": "Efficient Neural Audio Synthesis"
        }
      },
      {
        "id": "360287970189639836",
        "label": "Paper",
        "properties": {
          "title": "Gorilla: Large Language Model Connected with Massive APIs"
        }
      },
      {
        "id": "360287970189639889",
        "label": "Paper",
        "properties": {
          "title": "Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer"
        }
      },
      {
        "id": "360287970189639820",
        "label": "Paper",
        "properties": {
          "title": "EXAMS: A Multi-Subject High School Examinations Dataset for Cross-Lingual and Multilingual Question Answering"
        }
      },
      {
        "id": "360287970189639898",
        "label": "Paper",
        "properties": {
          "title": "Demystifying CLIP Data"
        }
      },
      {
        "id": "360287970189639899",
        "label": "Paper",
        "properties": {
          "title": "Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM"
        }
      },
      {
        "id": "360287970189639807",
        "label": "Paper",
        "properties": {
          "title": "Ring Attention with Blockwise Transformers for Near-Infinite Context"
        }
      },
      {
        "id": "360287970189639714",
        "label": "Paper",
        "properties": {
          "title": "Quantifying Memorization Across Neural Language Models"
        }
      },
      {
        "id": "360287970189639763",
        "label": "Paper",
        "properties": {
          "title": "MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models"
        }
      },
      {
        "id": "360287970189639904",
        "label": "Paper",
        "properties": {
          "title": "Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism"
        }
      },
      {
        "id": "360287970189639907",
        "label": "Paper",
        "properties": {
          "title": "AGIEval: A Human-Centric Benchmark for Evaluating Foundation Models"
        }
      }
    ]
  },
  {
    "category_id": 2,
    "name": "Natural Language Understanding and Reasoning",
    "description": "Experiments that test models' ability to understand and reason with natural language, including question answering, commonsense reasoning, and chain-of-thought prompting.",
    "children": [
      {
        "id": "360287970189639689",
        "label": "Paper",
        "properties": {
          "title": "WebGPT: Browser-assisted question-answering with human feedback"
        }
      },
      {
        "id": "360287970189639692",
        "label": "Paper",
        "properties": {
          "title": "Metacognitive Capabilities of LLMs: An Exploration in Mathematical Problem Solving"
        }
      },
      {
        "id": "360287970189639696",
        "label": "Paper",
        "properties": {
          "title": "TALM: Tool Augmented Language Models"
        }
      },
      {
        "id": "360287970189639699",
        "label": "Paper",
        "properties": {
          "title": "Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them"
        }
      },
      {
        "id": "360287970189639703",
        "label": "Paper",
        "properties": {
          "title": "Mind the instructions: a holistic evaluation of consistency and interactions in prompt-based learning"
        }
      },
      {
        "id": "360287970189639704",
        "label": "Paper",
        "properties": {
          "title": "Video-LLaVA: Learning United Visual Representation by Alignment Before Projection"
        }
      },
      {
        "id": "360287970189639708",
        "label": "Paper",
        "properties": {
          "title": "Common 7B Language Models Already Possess Strong Math Capabilities"
        }
      },
      {
        "id": "360287970189639709",
        "label": "Paper",
        "properties": {
          "title": "GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solvers"
        }
      },
      {
        "id": "360287970189639711",
        "label": "Paper",
        "properties": {
          "title": "Does your data spark joy? Performance gains from domain upsampling at the end of training"
        }
      },
      {
        "id": "360287970189639719",
        "label": "Paper",
        "properties": {
          "title": "Measuring Massive Multitask Language Understanding"
        }
      },
      {
        "id": "360287970189639728",
        "label": "Paper",
        "properties": {
          "title": "Automatically Correcting Large Language Models: Surveying the landscape of diverse self-correction strategies"
        }
      },
      {
        "id": "360287970189639729",
        "label": "Paper",
        "properties": {
          "title": "SocialIQA: Commonsense Reasoning about Social Interactions"
        }
      },
      {
        "id": "360287970189639716",
        "label": "Paper",
        "properties": {
          "title": "CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge"
        }
      },
      {
        "id": "360287970189639732",
        "label": "Paper",
        "properties": {
          "title": "Learning Transferable Visual Models From Natural Language Supervision"
        }
      },
      {
        "id": "360287970189639734",
        "label": "Paper",
        "properties": {
          "title": "QuAC : Question Answering in Context"
        }
      },
      {
        "id": "360287970189639739",
        "label": "Paper",
        "properties": {
          "title": "MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark"
        }
      },
      {
        "id": "360287970189639741",
        "label": "Paper",
        "properties": {
          "title": "Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference"
        }
      },
      {
        "id": "360287970189639743",
        "label": "Paper",
        "properties": {
          "title": "ZeroSCROLLS: A Zero-Shot Benchmark for Long Text Understanding"
        }
      },
      {
        "id": "360287970189639753",
        "label": "Paper",
        "properties": {
          "title": "Large Language Models Are Not Robust Multiple Choice Selectors"
        }
      },
      {
        "id": "360287970189639690",
        "label": "Paper",
        "properties": {
          "title": "TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension"
        }
      },
      {
        "id": "360287970189639740",
        "label": "Paper",
        "properties": {
          "title": "DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs"
        }
      },
      {
        "id": "360287970189639680",
        "label": "Paper",
        "properties": {
          "title": "Bag of Tricks for Efficient Text Classification"
        }
      },
      {
        "id": "360287970189639769",
        "label": "Paper",
        "properties": {
          "title": "Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding"
        }
      },
      {
        "id": "360287970189639694",
        "label": "Paper",
        "properties": {
          "title": "Monte Carlo Tree Search Boosts Reasoning via Iterative Preference Learning"
        }
      },
      {
        "id": "360287970189639713",
        "label": "Paper",
        "properties": {
          "title": "ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning"
        }
      },
      {
        "id": "360287970189639707",
        "label": "Paper",
        "properties": {
          "title": "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding"
        }
      },
      {
        "id": "360287970189639787",
        "label": "Paper",
        "properties": {
          "title": "ReAct: Synergizing Reasoning and Acting in Language Models"
        }
      },
      {
        "id": "360287970189639791",
        "label": "Paper",
        "properties": {
          "title": "CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data"
        }
      },
      {
        "id": "360287970189639792",
        "label": "Paper",
        "properties": {
          "title": "NExT-QA:Next Phase of Question-Answering to Explaining Temporal Actions"
        }
      },
      {
        "id": "360287970189639795",
        "label": "Paper",
        "properties": {
          "title": "CoVoST 2 and Massively Multilingual Speech-to-Text Translation"
        }
      },
      {
        "id": "360287970189639798",
        "label": "Paper",
        "properties": {
          "title": "MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalisms"
        }
      },
      {
        "id": "360287970189639735",
        "label": "Paper",
        "properties": {
          "title": "MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuning"
        }
      },
      {
        "id": "360287970189639809",
        "label": "Paper",
        "properties": {
          "title": "L-Eval: Instituting Standardized Evaluation for Long Context Language Models"
        }
      },
      {
        "id": "360287970189639810",
        "label": "Paper",
        "properties": {
          "title": "VQA: Visual Question Answering"
        }
      },
      {
        "id": "360287970189639811",
        "label": "Paper",
        "properties": {
          "title": "Learning From Mistakes Makes LLM Better Reasoner"
        }
      },
      {
        "id": "360287970189639812",
        "label": "Paper",
        "properties": {
          "title": "GPT-4 Technical Report"
        }
      },
      {
        "id": "360287970189639813",
        "label": "Paper",
        "properties": {
          "title": "Measuring Mathematical Problem Solving With the MATH Dataset"
        }
      },
      {
        "id": "360287970189639821",
        "label": "Paper",
        "properties": {
          "title": "Robust Speech Recognition via Large-Scale Weak Supervision"
        }
      },
      {
        "id": "360287970189639824",
        "label": "Paper",
        "properties": {
          "title": "LIMA: Less Is More for Alignment"
        }
      },
      {
        "id": "360287970189639826",
        "label": "Paper",
        "properties": {
          "title": "PIQA: Reasoning about Physical Commonsense in Natural Language"
        }
      },
      {
        "id": "360287970189639691",
        "label": "Paper",
        "properties": {
          "title": "Exploring the Limits of Weakly Supervised Pretraining"
        }
      },
      {
        "id": "360287970189639702",
        "label": "Paper",
        "properties": {
          "title": "Self-Refine: Iterative Refinement with Self-Feedback"
        }
      },
      {
        "id": "360287970189639799",
        "label": "Paper",
        "properties": {
          "title": "Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering"
        }
      },
      {
        "id": "360287970189639837",
        "label": "Paper",
        "properties": {
          "title": "STaR: Bootstrapping Reasoning With Reasoning"
        }
      },
      {
        "id": "360287970189639838",
        "label": "Paper",
        "properties": {
          "title": "ActivityNet-QA: A Dataset for Understanding Complex Web Videos via Question Answering"
        }
      },
      {
        "id": "360287970189639844",
        "label": "Paper",
        "properties": {
          "title": "Emergent Abilities of Large Language Models"
        }
      },
      {
        "id": "360287970189639845",
        "label": "Paper",
        "properties": {
          "title": "Reframing Instructional Prompts to GPTk's Language"
        }
      },
      {
        "id": "360287970189639849",
        "label": "Paper",
        "properties": {
          "title": "Illuminating search spaces by mapping elites"
        }
      },
      {
        "id": "360287970189639684",
        "label": "Paper",
        "properties": {
          "title": "Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity"
        }
      },
      {
        "id": "360287970189639853",
        "label": "Paper",
        "properties": {
          "title": "WorldSense: A Synthetic Benchmark for Grounded Reasoning in Large Language Models"
        }
      },
      {
        "id": "360287970189639832",
        "label": "Paper",
        "properties": {
          "title": "Magicoder: Empowering Code Generation with OSS-Instruct"
        }
      },
      {
        "id": "360287970189639778",
        "label": "Paper",
        "properties": {
          "title": "Efficient Estimation of Word Representations in Vector Space"
        }
      },
      {
        "id": "360287970189639744",
        "label": "Paper",
        "properties": {
          "title": "Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks"
        }
      },
      {
        "id": "360287970189639860",
        "label": "Paper",
        "properties": {
          "title": "LLaMA: Open and Efficient Foundation Language Models"
        }
      },
      {
        "id": "360287970189639869",
        "label": "Paper",
        "properties": {
          "title": "Scaling Instruction-Finetuned Language Models"
        }
      },
      {
        "id": "360287970189639871",
        "label": "Paper",
        "properties": {
          "title": "WinoGrande: An Adversarial Winograd Schema Challenge at Scale"
        }
      },
      {
        "id": "360287970189639872",
        "label": "Paper",
        "properties": {
          "title": "Toolformer: Language Models Can Teach Themselves to Use Tools"
        }
      },
      {
        "id": "360287970189639863",
        "label": "Paper",
        "properties": {
          "title": "Generating Sequences by Learning to Self-Correct"
        }
      },
      {
        "id": "360287970189639874",
        "label": "Paper",
        "properties": {
          "title": "ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solving"
        }
      },
      {
        "id": "360287970189639875",
        "label": "Paper",
        "properties": {
          "title": "Conic10K: A Challenging Math Problem Understanding and Reasoning Dataset"
        }
      },
      {
        "id": "360287970189639879",
        "label": "Paper",
        "properties": {
          "title": "TVQA: Localized, Compositional Video Question Answering"
        }
      },
      {
        "id": "360287970189639883",
        "label": "Paper",
        "properties": {
          "title": "Improved Baselines with Visual Instruction Tuning"
        }
      },
      {
        "id": "360287970189639885",
        "label": "Paper",
        "properties": {
          "title": "SQuAD: 100,000+ Questions for Machine Comprehension of Text"
        }
      },
      {
        "id": "360287970189639712",
        "label": "Paper",
        "properties": {
          "title": "Perception Test: A Diagnostic Benchmark for Multimodal Video Models"
        }
      },
      {
        "id": "360287970189639745",
        "label": "Paper",
        "properties": {
          "title": "Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer"
        }
      },
      {
        "id": "360287970189639896",
        "label": "Paper",
        "properties": {
          "title": "Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observations"
        }
      },
      {
        "id": "360287970189639700",
        "label": "Paper",
        "properties": {
          "title": "An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale"
        }
      },
      {
        "id": "360287970189639771",
        "label": "Paper",
        "properties": {
          "title": "Finetuned Language Models Are Zero-Shot Learners"
        }
      },
      {
        "id": "360287970189639710",
        "label": "Paper",
        "properties": {
          "title": "Training language models to follow instructions with human feedback"
        }
      },
      {
        "id": "360287970189639829",
        "label": "Paper",
        "properties": {
          "title": "DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models"
        }
      },
      {
        "id": "360287970189639909",
        "label": "Paper",
        "properties": {
          "title": "mPLUG-Owl: Modularization Empowers Large Language Models with Multimodality"
        }
      },
      {
        "id": "360287970189639698",
        "label": "Paper",
        "properties": {
          "title": "Calibrate Before Use: Improving Few-Shot Performance of Language Models"
        }
      },
      {
        "id": "360287970189639833",
        "label": "Paper",
        "properties": {
          "title": "Chain-of-Thought Prompting Elicits Reasoning in Large Language Models"
        }
      },
      {
        "id": "360287970189639759",
        "label": "Paper",
        "properties": {
          "title": "GAIA: a benchmark for General AI Assistants"
        }
      }
    ]
  },
  {
    "category_id": 3,
    "name": "Data Preprocessing and Filtering",
    "description": "Experiments that investigate the effects of preprocessing techniques and data filtering on model performance and generalization.",
    "children": [
      {
        "id": "360287970189639715",
        "label": "Paper",
        "properties": {
          "title": "Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks"
        }
      },
      {
        "id": "360287970189639752",
        "label": "Paper",
        "properties": {
          "title": "Towards VQA Models That Can Read"
        }
      },
      {
        "id": "360287970189639757",
        "label": "Paper",
        "properties": {
          "title": "Quantifying Variance in Evaluation Benchmarks"
        }
      },
      {
        "id": "360287970189639768",
        "label": "Paper",
        "properties": {
          "title": "A Diagram Is Worth A Dozen Images"
        }
      },
      {
        "id": "360287970189639779",
        "label": "Paper",
        "properties": {
          "title": "Extracting Training Data from Diffusion Models"
        }
      },
      {
        "id": "360287970189639781",
        "label": "Paper",
        "properties": {
          "title": "Billion-scale similarity search with GPUs"
        }
      },
      {
        "id": "360287970189639782",
        "label": "Paper",
        "properties": {
          "title": "OpenELM: An Efficient Language Model Family with Open Training and Inference Framework"
        }
      },
      {
        "id": "360287970189639783",
        "label": "Paper",
        "properties": {
          "title": "When Benchmarks are Targets: Revealing the Sensitivity of Large Language Model Leaderboards"
        }
      },
      {
        "id": "360287970189639786",
        "label": "Paper",
        "properties": {
          "title": "Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive"
        }
      },
      {
        "id": "360287970189639800",
        "label": "Paper",
        "properties": {
          "title": "Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale"
        }
      },
      {
        "id": "360287970189639803",
        "label": "Paper",
        "properties": {
          "title": "Training Deep Neural Networks with Joint Quantization and Pruning of Weights and Activations"
        }
      },
      {
        "id": "360287970189639797",
        "label": "Paper",
        "properties": {
          "title": "Training Verifiers to Solve Math Word Problems"
        }
      },
      {
        "id": "360287970189639831",
        "label": "Paper",
        "properties": {
          "title": "Efficient Memory Management for Large Language Model Serving with PagedAttention"
        }
      },
      {
        "id": "360287970189639841",
        "label": "Paper",
        "properties": {
          "title": "Deduplicating Training Data Makes Language Models Better"
        }
      },
      {
        "id": "360287970189639842",
        "label": "Paper",
        "properties": {
          "title": "Orca-Math: Unlocking the potential of SLMs in Grade School Math"
        }
      },
      {
        "id": "360287970189639843",
        "label": "Paper",
        "properties": {
          "title": "Effective Long-Context Scaling of Foundation Models"
        }
      },
      {
        "id": "360287970189639802",
        "label": "Paper",
        "properties": {
          "title": "The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions"
        }
      },
      {
        "id": "360287970189639846",
        "label": "Paper",
        "properties": {
          "title": "Dynabench: Rethinking Benchmarking in NLP"
        }
      },
      {
        "id": "360287970189639848",
        "label": "Paper",
        "properties": {
          "title": "Text-Free Prosody-Aware Generative Spoken Language Modeling"
        }
      },
      {
        "id": "360287970189639858",
        "label": "Paper",
        "properties": {
          "title": "Multi-Task Learning for Front-End Text Processing in TTS"
        }
      },
      {
        "id": "360287970189639856",
        "label": "Paper",
        "properties": {
          "title": "#InsTag: Instruction Tagging for Analyzing Supervised Fine-tuning of Large Language Models"
        }
      },
      {
        "id": "360287970189639822",
        "label": "Paper",
        "properties": {
          "title": "FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech"
        }
      },
      {
        "id": "360287970189639862",
        "label": "Paper",
        "properties": {
          "title": "What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning"
        }
      },
      {
        "id": "360287970189639897",
        "label": "Paper",
        "properties": {
          "title": "Emu Video: Factorizing Text-to-Video Generation by Explicit Image Conditioning"
        }
      },
      {
        "id": "360287970189639900",
        "label": "Paper",
        "properties": {
          "title": "RACE: Large-scale ReAding Comprehension Dataset From Examinations"
        }
      },
      {
        "id": "360287970189639902",
        "label": "Paper",
        "properties": {
          "title": "AudioChatLlama: Towards General-Purpose Speech Abilities for LLMs"
        }
      },
      {
        "id": "360287970189639892",
        "label": "Paper",
        "properties": {
          "title": "Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks"
        }
      },
      {
        "id": "360287970189639911",
        "label": "Paper",
        "properties": {
          "title": "PAWS: Paraphrase Adversaries from Word Scrambling"
        }
      },
      {
        "id": "360287970189639913",
        "label": "Paper",
        "properties": {
          "title": "A Self-Supervised Descriptor for Image Copy Detection"
        }
      },
      {
        "id": "360287970189639916",
        "label": "Paper",
        "properties": {
          "title": "Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations?"
        }
      }
    ]
  },
  {
    "category_id": 4,
    "name": "Tool Augmentation and Multimodal Learning",
    "description": "Experiments that explore how models can be augmented with external tools or learn from multiple modalities (e.g., text and images).",
    "children": [
      {
        "id": "360287970189639683",
        "label": "Paper",
        "properties": {
          "title": "Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V"
        }
      },
      {
        "id": "360287970189639685",
        "label": "Paper",
        "properties": {
          "title": "Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding"
        }
      },
      {
        "id": "360287970189639693",
        "label": "Paper",
        "properties": {
          "title": "Foundation Models for Video Understanding: A Survey"
        }
      },
      {
        "id": "360287970189639695",
        "label": "Paper",
        "properties": {
          "title": "Scaling Rectified Flow Transformers for High-Resolution Image Synthesis"
        }
      },
      {
        "id": "360287970189639717",
        "label": "Paper",
        "properties": {
          "title": "Perceiver: General Perception with Iterative Attention"
        }
      },
      {
        "id": "360287970189639731",
        "label": "Paper",
        "properties": {
          "title": "PromptTTS: Controllable Text-to-Speech with Text Descriptions"
        }
      },
      {
        "id": "360287970189639736",
        "label": "Paper",
        "properties": {
          "title": "Toward Joint Language Modeling for Speech Units and Text"
        }
      },
      {
        "id": "360287970189639756",
        "label": "Paper",
        "properties": {
          "title": "Spirit LM: Interleaved Spoken and Written Language Model"
        }
      },
      {
        "id": "360287970189639750",
        "label": "Paper",
        "properties": {
          "title": "ViP-LLaVA: Making Large Multimodal Models Understand Arbitrary Visual Prompts"
        }
      },
      {
        "id": "360287970189639793",
        "label": "Paper",
        "properties": {
          "title": "MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI"
        }
      },
      {
        "id": "360287970189639805",
        "label": "Paper",
        "properties": {
          "title": "WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct"
        }
      },
      {
        "id": "360287970189639806",
        "label": "Paper",
        "properties": {
          "title": "Gender Bias in Machine Translation"
        }
      },
      {
        "id": "360287970189639814",
        "label": "Paper",
        "properties": {
          "title": "ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection"
        }
      },
      {
        "id": "360287970189639816",
        "label": "Paper",
        "properties": {
          "title": "Iterative Reasoning Preference Optimization"
        }
      },
      {
        "id": "360287970189639823",
        "label": "Paper",
        "properties": {
          "title": "Know What You Don't Know: Unanswerable Questions for SQuAD"
        }
      },
      {
        "id": "360287970189639835",
        "label": "Paper",
        "properties": {
          "title": "Purple Llama CyberSecEval: A Secure Coding Benchmark for Language Models"
        }
      },
      {
        "id": "360287970189639850",
        "label": "Paper",
        "properties": {
          "title": "Learning Video Representations from Large Language Models"
        }
      },
      {
        "id": "360287970189639852",
        "label": "Paper",
        "properties": {
          "title": "Attention Is All You Need"
        }
      },
      {
        "id": "360287970189639855",
        "label": "Paper",
        "properties": {
          "title": "API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMs"
        }
      },
      {
        "id": "360287970189639857",
        "label": "Paper",
        "properties": {
          "title": "Bilateral Multi-Perspective Matching for Natural Language Sentences"
        }
      },
      {
        "id": "360287970189639861",
        "label": "Paper",
        "properties": {
          "title": "Verb Semantics and Lexical Selection"
        }
      },
      {
        "id": "360287970189639864",
        "label": "Paper",
        "properties": {
          "title": "MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action"
        }
      },
      {
        "id": "360287970189639751",
        "label": "Paper",
        "properties": {
          "title": "VioLA: Unified Codec Language Models for Speech Recognition, Synthesis, and Translation"
        }
      },
      {
        "id": "360287970189639886",
        "label": "Paper",
        "properties": {
          "title": "Conformer: Convolution-augmented Transformer for Speech Recognition"
        }
      },
      {
        "id": "360287970189639724",
        "label": "Paper",
        "properties": {
          "title": "Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models"
        }
      },
      {
        "id": "360287970189639905",
        "label": "Paper",
        "properties": {
          "title": "Crosslingual Generalization through Multitask Finetuning"
        }
      }
    ]
  },
  {
    "category_id": 5,
    "name": "Security and Safety Evaluations",
    "description": "Experiments that evaluate the security and safety behaviors of models, including their response to unsafe prompts and potential biases.",
    "children": [
      {
        "id": "360287970189639686",
        "label": "Paper",
        "properties": {
          "title": "Self-supervised Learning with Random-projection Quantizer for Speech Recognition"
        }
      },
      {
        "id": "360287970189639721",
        "label": "Paper",
        "properties": {
          "title": "TOOLVERIFIER: Generalization to New Tools via Self-Verification"
        }
      },
      {
        "id": "360287970189639723",
        "label": "Paper",
        "properties": {
          "title": "Jailbreaking Black Box Large Language Models in Twenty Queries"
        }
      },
      {
        "id": "360287970189639726",
        "label": "Paper",
        "properties": {
          "title": "Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts"
        }
      },
      {
        "id": "360287970189639730",
        "label": "Paper",
        "properties": {
          "title": "VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretation"
        }
      },
      {
        "id": "360287970189639758",
        "label": "Paper",
        "properties": {
          "title": "A Survey of Reinforcement Learning from Human Feedback"
        }
      },
      {
        "id": "360287970189639760",
        "label": "Paper",
        "properties": {
          "title": "MAmmoTH2: Scaling Instructions from the Web"
        }
      },
      {
        "id": "360287970189639776",
        "label": "Paper",
        "properties": {
          "title": "DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter"
        }
      },
      {
        "id": "360287970189639780",
        "label": "Paper",
        "properties": {
          "title": "Evaluation data contamination in LLMs: how do we measure it and (when) does it matter?"
        }
      },
      {
        "id": "360287970189639790",
        "label": "Paper",
        "properties": {
          "title": "Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations"
        }
      },
      {
        "id": "360287970189639755",
        "label": "Paper",
        "properties": {
          "title": "SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models"
        }
      },
      {
        "id": "360287970189639772",
        "label": "Paper",
        "properties": {
          "title": "Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation"
        }
      },
      {
        "id": "360287970189639867",
        "label": "Paper",
        "properties": {
          "title": "XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models"
        }
      },
      {
        "id": "360287970189639847",
        "label": "Paper",
        "properties": {
          "title": "Solving math word problems with process- and outcome-based feedback"
        }
      },
      {
        "id": "360287970189639870",
        "label": "Paper",
        "properties": {
          "title": "PAL: Program-aided Language Models"
        }
      },
      {
        "id": "360287970189639851",
        "label": "Paper",
        "properties": {
          "title": "Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity"
        }
      },
      {
        "id": "360287970189639880",
        "label": "Paper",
        "properties": {
          "title": "Leveraging Large Language Models for Multiple Choice Question Answering"
        }
      },
      {
        "id": "360287970189639882",
        "label": "Paper",
        "properties": {
          "title": "Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations"
        }
      },
      {
        "id": "360287970189639888",
        "label": "Paper",
        "properties": {
          "title": "PaLM: Scaling Language Modeling with Pathways"
        }
      },
      {
        "id": "360287970189639893",
        "label": "Paper",
        "properties": {
          "title": "Adversarial Examples for Evaluating Reading Comprehension Systems"
        }
      },
      {
        "id": "360287970189639877",
        "label": "Paper",
        "properties": {
          "title": "CyberSecEval 2: A Wide-Ranging Cybersecurity Evaluation Suite for Large Language Models"
        }
      },
      {
        "id": "360287970189639818",
        "label": "Paper",
        "properties": {
          "title": "Direct Preference Optimization: Your Language Model is Secretly a Reward Model"
        }
      },
      {
        "id": "360287970189639901",
        "label": "Paper",
        "properties": {
          "title": "Alpaca against Vicuna: Using LLMs to Uncover Memorization of LLMs"
        }
      },
      {
        "id": "360287970189639878",
        "label": "Paper",
        "properties": {
          "title": "Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge"
        }
      },
      {
        "id": "360287970189639706",
        "label": "Paper",
        "properties": {
          "title": "Gemma: Open Models Based on Gemini Research and Technology"
        }
      },
      {
        "id": "360287970189639915",
        "label": "Paper",
        "properties": {
          "title": "BASE Layers: Simplifying Training of Large, Sparse Models"
        }
      },
      {
        "id": "360287970189639840",
        "label": "Paper",
        "properties": {
          "title": "MuTox: Universal MUltilingual Audio-based TOXicity Dataset and Zero-shot Detector"
        }
      }
    ]
  }
]