Acknowledgments & Citation
Acknowledging TeXRA
We are thrilled if TeXRA proves useful for your academic research! While not required, if TeXRA played a significant role in your work—particularly if used as part of a study or evaluation involving LLMs or academic research tools—we would greatly appreciate an acknowledgment or citation if possible.
As the project evolves, we will provide a preferred citation format (e.g., a white paper or software citation). Please refer to the TeXRA GitHub repository or the texra.ai website for future citation details.
Your feedback and potential acknowledgments help support the continued development and improvement of TeXRA. Thank you for using it!
Supporting TeXRA
If TeXRA helps you publish faster, graduate sooner, or simply reduces your LaTeX-induced stress levels, consider supporting its development:
Conceptual Background & References
TeXRA's design draws inspiration from several key concepts in AI and software development:
- Agentic Workflows & Tool Use [1]: The core idea involves AI agents executing tasks augmented by specialized tools (e.g.,
texcount). This allows LLMs to leverage external capabilities for tasks requiring precision or specific knowledge beyond their training data. - Chain-of-Thought (CoT) Reasoning [2]: For complex agents, TeXRA employs techniques inspired by Chain-of-Thought prompting, encouraging models to "think step-by-step" (often visible in the
<scratchpad>sections of logs) before producing a final output. - Reflection & Action [3, 4]: The optional "Reflect" step, combined with the agent's ability to act (edit text, use tools), draws inspiration from frameworks like ReAct and Reflexion, allowing iterative refinement based on self-critique or environmental feedback.
- Structured Prompting (YAML + Jinja): The use of YAML for structure and Jinja for templating within prompts allows for complex logic, dynamic content injection, and better maintainability, drawing inspiration from approaches seen in libraries like Prompt Poet. The support for inheritance and modularity allows for a more flexible and reusable prompt design.
We believe combining these concepts provides a robust and adaptable platform for AI-powered academic writing assistance.
References
[1] Schick, T., Dwivedi-Yu, J., Dessì, R., Raileanu, R., Lomeli, M., Zettlemoyer, L., Cancedda, N., & Scialom, T. (2023). Toolformer: Language Models Can Teach Themselves to Use Tools. arXiv preprint arXiv:2302.04761.
[2] Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E., Le, Q., & Zhou, D. (2022). Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. Advances in Neural Information Processing Systems (NeurIPS), 35, 24824–24837.
[3] Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., & Cao, Y. (2023). ReAct: Synergizing Reasoning and Acting in Language Models. International Conference on Learning Representations (ICLR).
[4] Shinn, N., Cassano, F., Gopinath, A., Narasimhan, K., & Yao, S. (2023). Reflexion: Language Agents with Verbal Reinforcement Learning. In Advances in Neural Information Processing Systems 36 (NeurIPS 2023).