LLM papers
-
Direct Preference Optimization: Your Language Model is Secretly a Reward Model 논문 리뷰LLM papers 2024. 5. 22. 16:17
[2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Model (arxiv.org) Direct Preference Optimization: Your Language Model is Secretly a Reward ModelWhile large-scale unsupervised language models (LMs) learn broad world knowledge and some reasoning skills, achieving precise control of their behavior is difficult due to the completely unsupervised nature of their ..
-
DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models 논문 리뷰LLM papers 2024. 5. 20. 11:29
[2306.11698] DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models (arxiv.org) DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT ModelsGenerative Pre-trained Transformer (GPT) models have exhibited exciting progress in their capabilities, capturing the interest of practitioners and the public alike. Yet, while the literature on the trustworthiness of GPT mo..
-
Scaling Data-Constrained Language Models 논문 리뷰LLM papers 2024. 5. 11. 15:12
[2305.16264] Scaling Data-Constrained Language Models (arxiv.org) Scaling Data-Constrained Language ModelsThe current trend of scaling language models involves increasing both parameter count and training dataset size. Extrapolating this trend suggests that training dataset size may soon be limited by the amount of text data available on the internet. Motivatearxiv.org AbstractLanguage Model을 스케..
-
QLoRA: Efficient Finetuning of Quantized LLMs 논문 리뷰LLM papers 2024. 5. 4. 17:03
[2305.14314] QLoRA: Efficient Finetuning of Quantized LLMs (arxiv.org) QLoRA: Efficient Finetuning of Quantized LLMsWe present QLoRA, an efficient finetuning approach that reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU while preserving full 16-bit finetuning task performance. QLoRA backpropagates gradients through a frozen, 4-bit quanarxiv.org AbstractQLoRA- 6..
-
ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings 논문 리뷰LLM papers 2024. 4. 28. 17:18
[2305.11554] ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings (arxiv.org) ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool EmbeddingsAugmenting large language models (LLMs) with external tools has emerged as a promising approach to solving complex problems. However, traditional methods, which finetune LLMs with tool demonstration data,..
-
Toolformer: Language Models Can Teach Themselves to Use Tools 논문 리뷰LLM papers 2024. 4. 2. 15:50
[2302.04761] Toolformer: Language Models Can Teach Themselves to Use Tools (arxiv.org) Toolformer: Language Models Can Teach Themselves to Use Tools Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where muc..
-
Jailbroken: How Does LLM Safety Training Fail? 논문 리뷰LLM papers 2024. 3. 27. 22:09
[2307.02483] Jailbroken: How Does LLM Safety Training Fail? (arxiv.org) Jailbroken: How Does LLM Safety Training Fail? Large language models trained for safety and harmlessness remain susceptible to adversarial misuse, as evidenced by the prevalence of "jailbreak" attacks on early releases of ChatGPT that elicit undesired behavior. Going beyond recognition of the issue, we arxiv.org Abstract LLM..
-
Are Emergent Abilities of Large Language Models a Mirage?LLM papers 2024. 3. 20. 22:49
[2304.15004] Are Emergent Abilities of Large Language Models a Mirage? (arxiv.org) Are Emergent Abilities of Large Language Models a Mirage? Recent work claims that large language models display emergent abilities, abilities not present in smaller-scale models that are present in larger-scale models. What makes emergent abilities intriguing is two-fold: their sharpness, transitioning seemingly a..