GPT-4 seems to have become less powerful recently, with features like omitting parts of replies, shortening responses, and even cutting corners when replying with code. Two solutions are shared: using the official GPTs to experience the classic smoothness before November, or using a set of super combination commands to enhance performance. Denis Shiryaev also suggests adding the command "- today is May (Not a December)" to counter GPT-4's laziness in December.
This article introduces the functionality and usage of the "Triple Exponential Moving Average (TEMA) Channel" as a trend indicator. The TEMA indicator smooths short-term fluctuations by using multiple moving averages and adjusts for lag in the formula. It is best suited for long-term trend trading and can be used in combination with other oscillators or technical indicators. The article also provides a script example using TEMA to display channels of price support and resistance.
This article introduces the Klinger Oscillator, which is a both mysterious and practical technical indicator created by Stephen Klinger. The indicator can help determine the strength of market trends and capture short-term fluctuations and rises and falls in the market. The article also provides the source code for the Keltner Channel in the Tongda Xing software, which is used to construct the Keltner Channel.
This article introduces the "Know Sure Thing Indicator" (KST), which is a magical technical indicator that calculates a mysterious value by observing the rate of change (ROC) and the combination of simple moving averages (SMA) over four periods. KST behaves differently in analyzing bull markets and bear markets, and it has the concepts of overbought and oversold. The article also provides a Pine Script code example for plotting the KST indicator.
This post discusses the performance of GPT-4-128K with long context recall. The findings reveal that recall performance starts to degrade above 73K tokens, low recall is correlated with facts placed between 7%-50% document depth, and facts placed at the beginning or 2nd half of the document are recalled better. It is advised not to guarantee fact retrieval, reduce context for more accuracy, and consider the position of facts. The process involved using Paul Graham essays as background tokens and evaluating GPT-4's answers. Further steps include using a sigmoid distribution and key:value retrieval. More testing is needed to fully understand GPT4's abilities.
This post discusses the performance of Claude 2.1, an LLM model, in recalling facts at different document depths. The findings indicate that facts at the top and bottom of the document were recalled with high accuracy, while performance decreased towards the middle. It is suggested to experiment with prompts and conduct A/B tests to improve retrieval accuracy, not to assume guaranteed retrieval of facts, reduce context length for better accuracy, and consider the position of facts within the document. The test aimed to gain insights into LLM performance and transfer that knowledge to practical use cases.
This article introduces a limit testing on large models, which significantly improves the performance of GPT-4 and Claude2.1 by adding specific prompt statements at the beginning of the responses. The test results show that large models have difficulties in finding specific sentences, but this method can address the issue. In addition, the Kimi team from the Dark Side of the Moon also proposes different solutions and achieves good results. The entire experiment demonstrates that the performance of large models is subject to certain limitations, but it can be improved by appropriate prompts and adjustments.
This article introduces the principles and applications of the Keltner Channel as a technical analysis tool. The Keltner Channel combines the analysis of price volatility and trading volume to identify market trends and turning points. The article also provides an improved version of the code on the TongdaXin platform, making this indicator more accurate and useful in stock market technical analysis.
Historical volatility is an indicator that measures the price fluctuations in the stock market. It reveals the true pulse of the market by calculating the average deviation between financial instruments and their average prices. Hull Moving Average (HMA) is a fast and smooth technical indicator that provides a smooth and efficient analysis of market dynamics through a special calculation method. Understanding and using HMA can help traders capture subtle changes in market trends and select appropriate trading opportunities.
Historical volatility is an indicator that measures the price fluctuations in the stock market. It reveals the true pulse of the market by calculating the average deviation between financial instruments and their average prices. High volatility implies high-risk and high-return opportunities, while low volatility suggests relatively fewer profit opportunities. Traders can find their suitable volatility levels by comparing the volatilities of different securities and utilizing other technical analysis tools.
The document introduces the Fisher Transform, an indicator used in market analysis to capture price reversals. It explains the background and working principle of the Fisher Transform, as well as its strengths and weaknesses. The document also provides advanced techniques for using the indicator and includes the code implementation in Pine Script language.
Moving Average Envelope (ENV) is a simple yet powerful technical indicator that can help us gain insights into market trends. It consists of a baseline and two envelope lines, which are determined by setting fixed percentages. When the price breaks through the envelope lines, it is a significant signal that may indicate important price fluctuations. ENV can assist us in identifying overbought and oversold opportunities, but it should be used in conjunction with other technical tools. This indicator can provide us with a strong perspective to help us navigate the stock market with stability.