Loading
Life sciences · Preprint
arXiv · September 4, 2026
Raises a question worth testing. It does not answer one.
This preprint describes PRICE, a systems approach combining fine-tuning, prompting, and decoding strategies applied to a quantized LLaMA-3 model for short-term Bitcoin price forecasting. Ablation studies show each component contributes to accuracy and PRICE outperforms eight baseline models on validation and test sets. The work is methodologically sound as a technical systems study but lacks peer review, real-world trading validation, and independent replication; it remains exploratory and demonstrates engineering feasibility rather than clinical or investment utility.
Ablation study of LLM adaptation strategies for time-series forecasting. Bitcoin price historical time series; no human subjects.. Intervention: PRICE: combined adaptation of 4-bit quantized LLaMA-3 8B model via LoRA fine-tuning, recursive inference, integer-rounded values, CTF prompting, and zero-temperature decoding.. Compared with: Eight transformer-based and time-series foundation models..
LoRA enables efficient training on limited hardware and contributes to forecasting accuracy. Recursive multi-step inference improves accuracy compared to single-step forecasting. Integer-rounded numerical representation reduces forecasting errors versus other representations.
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
This is a systems engineering study of LLM adaptation choices for cryptocurrency forecasting using ablation methods; it demonstrates technical feasibility but lacks clinical or real-world validation, peer review, and independent replication.
As stated by the source record.
Quoted from the source exactly as published.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
Cryptocurrency markets exhibit extreme volatility and non-stationary dynamics that challenge conventional forecasting methods. Although Large Language Models (LLMs) have shown promise for time series forecasting, the combined effects of adaptation choices remain largely unexplored in financial settings. This study introduces PRICE, a structured approach for adapting LLMs to short-term Bitcoin price forecasting. Built on a 4-bit quantized LLaMA-3 8B model, PRICE investigates how fine-tuning, numerical representation, prompting, inference, and decoding jointly influence forecasting performance. PRICE integrates Parameter-efficient fine-tuning with Low-Rank Adaptation (LoRA), Recursive multi-step inference, Integer-rounded numerical representation, Context-Task-Format (CTF) prompting, and Exact zero-temperature decoding. Ablation studies show that each component contributes to forecasting accuracy and reliability. LoRA enables efficient training on limited hardware, recursive inference improves accuracy, integer-rounded values reduce errors, CTF prompting outperforms Chain-of-Thought, Implicit Chain-of-Thought (iCoT), and few-shot prompting, and zero-temperature decoding improves stability during recursive forecasting. Comparative evaluation against eight transformer-based and time-series foundation models shows that PRICE achieves the lowest forecasting errors on both validation and test sets while maintaining robust performance across evaluation periods. Despite being based on a model primarily pretrained on text rather than time-series data, PRICE achieves competitive or superior performance relative to specialized foundation models. These findings demonstrate that adaptation choices critically determine the accuracy and robustness of LLMs for numerical time-series forecasting.
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.