activity
20242026
most citedStochastic CHAOS: Why Deterministic Inference Kills, and Distributional Variability Is the Heartbeat of Artifical Cognition

1 citations · 1 across the 22 of their papers we have counts for

collaborators
Showing cs.LGShow all

11 papers · 1 filter

cs.LG2026

SPHERICAL KV: Angle-Domain Attention and Rate-Distortion Retention for Efficient Long-Context Inference

Anay Chauhan, Gurucharan Marthi Krishna Kumar, Arion Das +4

Long-context inference is increasingly constrained by the KV cache: resident memory grows with context length, and decoding becomes limited by repeated High Bandwidth Memory (HBM)…

cs.LG2026

MAAT: Multi-phase Adapter-Aware Targeted Unlearning

Suryash Yagnik, Shubham Gaur, Saksham Thakur +3

Machine unlearning evaluation is structurally skewed: Why-type questions, which probe causal and relational knowledge, comprise less than 0.06% of CounterFact, 0.6% of ZSRE, and le…

cs.LG2026

Towards Explainability of SLMs by investigating Token Level Activation

Sayantani Ghosh, Rajashik Datta, Amit Kumar Das +1

Transformer-based language models such as BERT having 110M+ parameters have revolutionized natural language understanding, yet their internal mechanisms remain largely opaque to re…

cs.LG2026

A New Technique for AI Explainability using Feature Association Map

Sayantani Ghosh, Amit Kumar Das, Amlan Chakrabarti

Lack of transparency in AI systems poses challenges in critical real-life applications. It is important to be able to explain the decisions of an AI system to ensure trust on the s…

cs.LG2026

PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training

Harsh Kumar, Rahul Maity, Tanmay Joshi +4

Aligned large language models (LLMs) remain vulnerable to adversarial manipulation, and their reliance on web-scale pretraining creates a subtle but consequential attack surface. W…

cs.LG2026

SPINAL -- Scaling-law and Preference Integration in Neural Alignment Layers

Arion Das, Partha Pratim Saha, Amit Dhanda +3

Direct Preference Optimization (DPO) is a principled, scalable alternative to RLHF for aligning large language models from pairwise preferences, but its internal geometric footprin…