Numerical Reasoning for Financial Reports
On the map: LLM reasoning → Reasoning for agents, planning and code
Abstract (arXiv)
Financial reports offer critical insights into a company's operations, yet their extensive length typically spanning 30 40 pages poses challenges for swift decision making in dynamic markets. To address this, we leveraged finetuned Large Language Models (LLMs) to distill key indicators and operational metrics from these reports basis questions from the user. We devised a method to locate critical data, and leverage the FinQA dataset to fine-tune both Llama-2 7B and T5 models for customized question answering. We achieved results comparable to baseline on the final numerical answer, a competitive accuracy in numerical reasoning and calculation.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 11 of its 16 references that are in the map. First posted on arXiv 22 Dec 2023.
Builds on
11 of 16 references.
References · 11
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
- 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,418
- 2022Scaling Instruction-Finetuned Language Modelscited by 3,024
- 2019DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphscited by 903
- 2022PAL: Program-aided Language Modelscited by 652
- 2022TabLLM: Few-shot Classification of Tabular Data with Large Language Modelscited by 277
- 2020HybridQA: A Dataset of Multi-Hop Question Answering over Tabular and Textual Datacited by 264
- 2023Instruction Tuned Models are Quick Learnerscited by 21
- 2022"John is 50 years old, can his son be 65?" Evaluating NLP Models' Understanding of Feasibilitycited by 15
- 2021Context-NER : Contextual Phrase Generation at Scalecited by 11
- 2021Transformer-Based Models for Question Answering on COVID19cited by 7
Built on by
Citing papers per year: 2024: 1; 2025: 2.
3 papers.
- 2024A Survey of Large Language Models for Financial Applications: Progress, Prospects and Challengescited by 90
- 2025Replay Failures as Successes: Sample-Efficient Reinforcement Learning for Instruction Followingcited by 8
- 2025Meeseeks: A Feedback-Driven, Iterative Self-Correction Benchmark evaluating LLMs' Instruction Following Capabilitycited by 3
- 2025Replay Failures as Successes: Sample-Efficient Reinforcement Learning for Instruction Following29 Dec 2025
- 2025Meeseeks: A Feedback-Driven, Iterative Self-Correction Benchmark evaluating LLMs' Instruction Following Capability30 Apr 2025
- 2024A Survey of Large Language Models for Financial Applications: Progress, Prospects and Challenges15 Jun 2024
Similar, not cited
Close in wording, no citation either way
- 2024FinLLMs: A Framework for Financial Reasoning Dataset Generation with Large Language Models
- 2026Document-Level Numerical Reasoning across Single and Multiple Tables in Financial Reports
- 2023BizBench: A Quantitative Reasoning Benchmark for Business and Finance
- 2026From Documents to Reasoning: A Validated Synthetic Data Pipeline and Semantic-Aware Fine-Tuning for Financial Numerical Reasoning
- 2024Evaluating LLMs' Mathematical Reasoning in Financial Document Question Answering
- 2025FinLFQA: Evaluating Attributed Text Generation of LLMs in Financial Long-Form Question Answering
- 2025Evaluating Large Language Models (LLMs) in Financial NLP: A Comparative Study on Financial Report Analysis
- 2026Structure First, Reason Next: Enhancing a Large Language Model using Knowledge Graph for Numerical Reasoning in Financial Documents