Research

Peer-reviewed research

Six papers across LLM interpretability, model compression, and AI-generated text detection, with two as first author. Each entry links to the paper or code.

First author
2026
First author

CircuitTier: Multi-Signal Interpretability for Efficient Model Compression

M. Srikar Vardhan, et al.
ACL Rolling Review 2026 · meta-review recommendation: Findings · committing to the IJCNLP-AACL cycle

A task-aware way to quantize large language models. Instead of compressing every component equally, CircuitTier fuses six interpretability signals to set per-component bit-widths, protecting the parts that matter for the task. On Gemma-2-2B for SQL generation it keeps about 99% of task accuracy at 3.84x compression, and the result holds across architectures, with 3.84x at 97.6% retention on Llama-3-8B.

2026
First author

Disentangling Direction and Magnitude in Transformer Representations

M. Srikar Vardhan, L. S. Teja
arXiv preprint, 2026 · A double dissociation through L2-matched perturbation analysis

An interpretability study that separates the direction and the magnitude of Transformer representations. Using L2-matched perturbations, it demonstrates a double dissociation between the two, clarifying which part actually carries task information inside the model.

Co-author
2025
Co-author

Fine-Grained Detection of AI-Generated Text Using Sentence-Level Segmentation

L. D. M. S. S. Teja, A. Yadagiri, P. Pakray, C. Chunka, M. S. Vardhan
Findings of IJCNLP-AACL 2025

Detecting AI-generated text at the sentence level rather than labeling a whole document, which allows much more precise attribution of which parts of a text were machine-written.

2025
Co-author

CNLP-NITS-PP at GenAI Detection Task 1: AI-Generated Text Using Transformer-Based Approaches

A. Yadagiri, S. T. Lekkala, M. S. Vardhan, P. Pakray, R. M. Krishna
1st Workshop on GenAI Content Detection (GenAIDetect), COLING 2025

A transformer-based system for the shared task on detecting machine-generated text.

2025
Co-author

nits_teja_srikar at GenAI Detection Task 2: Distinguishing Human and AI-Generated Essays

S. T. Lekkala, A. Yadagiri, M. S. Vardhan, P. Pakray
1st Workshop on GenAI Content Detection (GenAIDetect), COLING 2025

Machine learning and transformer models to tell human-written essays apart from AI-generated ones.

2025
Co-author

CNLP-NITS-PP at GenAI Detection Task 3: Cross-Domain Machine-Generated Text Detection Using DistilBERT

S. T. Lekkala, A. Yadagiri, M. S. Vardhan, P. Pakray
1st Workshop on GenAI Content Detection (GenAIDetect), COLING 2025

A DistilBERT-based approach for detecting machine-generated text that generalizes across domains.

Google Scholar ACL Anthology profile