Generate a large dataset of labeled sentences for training a FastText model to classify two intents: lab_guide_question and virtual_lab_manipulation. Each labeled sentence should represent one of these intents and must vary in structure, vocabulary, and style to ensure robustness in the training data. Follow these guidelines for each intent type: ### **1. lab_guide_question (50%)** These sentences should reflect questions typically asked by students about lab guides, scientific concepts, or theoretical aspects of an experiment. Include various phrasings, question structures, and scientific terminology. Example topics include: - Chemical properties (e.g., mass of mercury, atomic behavior) - Experimental concepts (e.g., inelastic collisions, reaction times) - Requirements for lab conditions (e.g., temperature settings, pressure levels) Examples: - "What is the atomic weight of mercury?" - "Why do mercury atoms show inelastic collisions?" - "What temperature should the reaction vessel be?" ### **2. virtual_lab_manipulation (50%)** These sentences should reflect actions or commands that users might give when interacting with a virtual lab simulation. Focus on practical tasks like adjusting settings, observing measurements, or controlling instruments. Vary the style between formal commands and casual inquiries. Examples: - "Set the oven to 300 degrees." - "Check the temperature of the furnace." - "Increase the reaction chamber's pressure." - "Find the VVR measurement." ### **Instructions:** 1. Ensure that each intent is clearly distinguishable by context and phrasing. 2. Include a wide range of variations for each example (e.g., formal vs. informal, direct vs. indirect questions). 3. Avoid using overly repetitive phrases; maintain natural language diversity. 4. Label each sentence with the appropriate intent in the format: __label__intent sentence. Example Outputs: __label__lab_guide_question What is the mass of mercury? __label__virtual_lab_manipulation Find the temperature setting for the oven. __label__lab_guide_question Why do we observe inelastic collisions in mercury atoms? __label__virtual_lab_manipulation Check the VVR display. You should generate NOTHING ELSE except the sample data in the example format! --- ilg.physics.ucsb.edu UCSB Physics Website and Manuals designed by Kelly Ann Pawlak, UCSB ILG (c) 2020 25–31 minutes The Franck-Hertz Experiment Part I: Introduction When quantum mechanics was first proposed, some physicists wondered if it wasn't just a "trick of the light" because all the phenomena that required quantization for their explanation (i.e., atomic spectra and the photoelectric effect) involved light. That changed when James Franck and Gustav Hertz provided evidence for quantum mechanics that did not involve light. In this lab you will carry out a modern version of the Nobel-prize winning Franck-Hertz experiment. Part II: Background History In 1911, James Franck was a 29-year old professor of Physics. Gustav Hertz was a 24-year old gradaute student. Franck and Hertz didn't set out to validate quantum theory. They conceived their experiment well before Niels Bohr proposed his model of the atom. Their goal was to better understand electron conductivity through gases. But by the time they built their experiment, collected their data and reported their results, it was 1914. The Bohr model of the atom was about one year old. The Franck-Hertz data offered unequivocal evidence that atoms can accept energy only in discrete (i.e., quantized) amounts. It was crucial to the acceptance of quantum theory because, at that time, photons were still somewhat mysterious but electrons were unquestionably real particles whose kinetic energy could be readily determined. This first-hand account of the historical context surrounding the Franck-Hertz experiment from Gustav Hertz himself is in German, but you can display English subtitles: Theory The Franck-Hertz experiment generates free electrons by heating a cathode inside an evacuated tube (i.e., a vacuum tube). The newly freed electrons accelerate towards an anode because a voltage difference, Va, is imposed between the anode and the cathode. As they move from cathode to anode, the electrons gain kinetic energy eVa, unless they run into something along the way and undergo an inelastic collision. "But what could the electrons run into?" you ask, "Didn't you just say they are in an evacuated tube?" Yes, the tube is evacuated, but it is not empty. There is a a small amount of mercury in the tube. Mercury is a liquid at room temperature and pressure. Under vacuum and with a little heat, it becomes a gas. When the tube is heated (in an oven), a low pressure gas of mercury atoms occupies the region between the cathode and anode. So the moving electrons can collide with, and scatter off of, the mercury atoms. Because the mass of a mercury atom (3×10−22g) is hundreds of thousands of times greater than the mass of an electron (9×10−28g), and because the kinetic energy of a mercury atom at 200∘C (∼0.04 eV) is much, much less than that of an electron that has traveled only a fraction of a millimeter toward the anode, if the two undergo an elastic collision, the electron's kinetic energy will barely change. If they undergo an inelastic collision, however, the electron could come away with much less kinetic energy. If Hg atoms could absorb an arbitrary amount of energy, electrons that collide with them would lose energy no matter what the value of Va. But, if Hg atoms can only absorb discrete energies from the electrons, the electrons that collide with them will only lose energy if they have kinetic energy greater than or equal to the smallest amount of energy that the mercury atom can absorb. The Franck-Hertz experiment involves changing Va and measuring the kinetic energy of the electrons that reach the anode. In theory, a sudden drop should occur when eVa becomes equal to the smallest amount of energy that a mercury atom can accept, an amount known as "the first excitation energy" of mercury. Let's call the accelerating voltage that satisfies this condition V1. When Va=V1, electrons will only have enough energy to excite mercury once they reach the anode. As Va is increased above V1, electrons will be able to excite mercury further from the anode and so continue to accelerate towards the anode after a collision. When Va>2eV1, it becomes possible for an electron to undergo two inelastic collisions, and again be left with little energy when it reaches the anode. This continues as the voltage is increased further: the average kinetic energy of electrons at the anode drops each time Va exceeds an integer multiple of mercury's first excitation energy.] But, you ask,what if V1 Generate a large dataset of labeled sentences for training a FastText model to classify two intents: lab_guide_question and virtual_lab_manipulation. Each labeled sentence should represent one of these intents and must vary in structure, vocabulary, and style to ensure robustness in the training data. Follow these guidelines for each intent type: ### **1. lab_guide_question (50%)** These sentences should reflect questions typically asked by students about lab guides, scientific concepts, or theoretical aspects of an experiment. Include various phrasings, question structures, and scientific terminology. They should relate to ideas or objects mentioned in the topics mentioned in the above article. - "What is the atomic weight of mercury?" - "Why do mercury atoms show inelastic collisions?" - "What temperature should the reaction vessel be?" ### **2. virtual_lab_manipulation (50%)** These sentences should reflect actions or commands that users might give when interacting with a virtual lab simulation. Focus on practical tasks like adjusting settings, observing measurements, or controlling instruments. Vary the style between formal commands and casual inquiries. - "Set the oven to 300 degrees." - "Check the temperature of the furnace." - "Increase the reaction chamber's pressure." - "Find the VVR measurement." 1. Ensure that each intent is clearly distinguishable by context and phrasing. 2. Include a wide range of variations for each example (e.g., formal vs. informal, direct vs. indirect questions, proper grammar vs improper slang). 3. Avoid using overly repetitive phrases; maintain natural language diversity. 4. Label each sentence with the appropriate intent in the format: __label__intent sentence. __label__lab_guide_question What is the mass of mercury? __label__virtual_lab_manipulation Find the temperature setting for the oven. __label__lab_guide_question Why do we observe inelastic collisions in mercury atoms? __label__virtual_lab_manipulation Check the VVR display. __label__lab_guide_question how to find thermometer? __label__virtual_lab_manipulation go to oven controls You should generate NOTHING ELSE except the sample data in the example format!