Papers by Rahul Nair
HalluDetect: Detecting, Mitigating, and Benchmarking Hallucinations in Conversational Systems in the Legal Domain (2025.emnlp-industry)
Copied to clipboard
Spandan Anaokar, Shrey Ganatra, Swapnil Bhattacharyya, Harshvivek Kashid, Shruthi N Nair, Reshma Sekhar, Siddharth Manohar, Rahul Hemrajani, Pushpak Bhattacharyya
| Challenge: | Large Language Models (LLMs) are widely used in industry but still produce hallucinations, limiting their reliability in critical applications. |
| Approach: | They propose to reduce hallucinations in consumer grievance chatbots by reducing their token accuracy by 0.4159 per turn. |
| Outcome: | The proposed system achieves an F1 score of 68.92% outperforming baseline detectors by 22.47% while maintaining the highest token accuracy. |
Towards Automated Extraction of Business Constraints from Unstructured Regulatory Text (C18-2)
Copied to clipboard
| Challenge: | a system for machine-driven annotations of legal documents is currently undergoing user trials within our organization. |
| Approach: | a system for machine-driven annotations of legal documents is presented . the system is currently undergoing user trials within our organization. |
| Outcome: | the proposed system is currently undergoing user trials within our organization. |
Ranking Large Language Models without Ground Truth (2024.findings-acl)
Copied to clipboard
| Challenge: | Evaluation and ranking of large language models has become a problem with the proliferation of these models and their impact. |
| Approach: | They propose to rank large language models without access to ground truth or reference responses . they propose to use triplets of models to evaluate the other two, correctly identifying the worst model in the triplet with high probability. |
| Outcome: | The proposed method reliably recovers true rankings without reference data on generative tasks. |