1 大模型文化道德

• Knowledge of cultural moral norms in large language models
  • url：https://aclanthology.org/2023.acl-long.26/
  • code：https://github.com/AidaRamezani/cultural_inference
• Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models
  • url：https://aclanthology.org/2023.acl-long.84/
  • code：https://github.com/myracheng/markedpersonas

2 长文本推理

• Open-ended Long Text Generation via Masked Language Modeling
  • url：https://aclanthology.org/2023.acl-long.13/
  • code：https://github.com/dropreg/OpenLTG-MLM
• Efficient Streaming Language Models with Attention Sinks
  • url：https://aclanthology.org/2023.acl-long.13/
  • code：https://github.com/mit-han-lab/streaming-llm

3 大模型推理

• In-Context Analogical Reasoning with Pre-Trained Language Models
  • url：https://aclanthology.org/2023.acl-long.109/
  • code：https://github.com/hxiaoyang/lm-raven
• Semantic-Oriented Unlabeled Priming for Large-Scale Language Models
  • url：https://aclanthology.org/2023.sustainlp-1.2/

4 大模型测评

• A Causal Framework to Quantify the Robustness of Mathematical Reasoning with Language Models
  • url：https://aclanthology.org/2023.acl-long.32/
  • code：https://github.com/alestolfo/causal-math
• Do language models have coherent mental models of everyday things?
  • url：https://aclanthology.org/2023.acl-long.106/
  • code：https://github.com/allenai/everyday-things/
• We Understand Elliptical Sentences, and Language Models should Too: A New Dataset for Studying Ellipsis and its Interaction with Thematic Fit
  • url：https://aclanthology.org/2023.acl-long.188/
  • code：https://github.com/Caput97/ELLie-ellipsis_and_thematic_fit_with_LMs
• Can ChatGPT Understand Causal Language in Science Claims?
  • url：https://aclanthology.org/2023.wassa-1.33/
  • code：null
• ChatGPT is fun, but it is not funny! Humor is still challenging Large Language Models
  • url：https://aclanthology.org/2023.wassa-1.29/
  • code：https://github.com/DLR-SC/JokeGPT-WASSA23
• Examining the Causal Impact of First Names on Language Models: The Case of Social Commonsense Reasoning
  • url：https://aclanthology.org/2023.trustnlp-1.7/
  • code：https://github.com/sullamij/Causal-First-Names/

5 数据生成

• Self-Instruct: Aligning LM with Self Generated Instructions
  • url：https://aclanthology.org/2023.acl-long.754/
  • code：https://github.com/yizhongw/self-instruct
• Increasing Diversity While Maintaining Accuracy: Text Data Generation with Large Language Models and Human Interventions
  • url：https://aclanthology.org/2023.acl-long.34/
  • code：null
• Instruction Induction: From Few Examples to Natural Language Task Descriptions
  • url：https://aclanthology.org/2023.acl-long.108/
  • code：https://github.com/orhonovich/instruction-induction
• How to Unleash the Power of Large Language Models for Few-shot Relation Extraction?
  • url：https://aclanthology.org/2023.sustainlp-1.13/
  • code：https://github.com/zjunlp/DeepKE/tree/main/example/llm

6 大模型微调

• Fine-tuning Happens in Tiny Subspaces: Exploring Intrinsic Task-specific Subspaces of Pre-trained Language Models
  • url：https://aclanthology.org/2023.acl-long.95/
  • code：null
• ADEPT: Adapter-based Efficient Prompt Tuning Approach for Language Models
  • url：https://aclanthology.org/2023.sustainlp-1.8/
  • code：https://github.com/Aditya-shahh/ADEPT

7 大模型与安全

• GPTs Don’t Keep Secrets: Searching for Backdoor Watermark Triggers in Autoregressive Language Models
  • url：https://aclanthology.org/2023.trustnlp-1.21/
  • code：https://github.com/evan-person/findingBackdoorWatermarks

8 水文

• Decoding Symbolism in Language Models
  • url：https://aclanthology.org/2023.acl-long.186/
  • code：https://github.com/MeiqiGuo/ACL2023-SymbA
• What is Wrong with Language Models that Can Not Tell a Story?
  • url：https://aclanthology.org/2023.wnu-1.8/
  • code：null
• YNU-HPCC at WASSA-2023 Shared Task 1: Large-scale Language Model with LoRA Fine-Tuning for Empathy Detection and Emotion Classification
  • url：https://aclanthology.org/2023.wassa-1.45/
  • code：null