collaborators
Showing cs.CLShow all

6 papers · 1 filter

cs.CL2026

Transactional Attention: Semantic Sponsorship for KV-Cache Retention

Abhinaba Basu

At K=16 tokens (0.4% of a 4K context), every existing KV-cache compression method achieves 0% on credential retrieval. The failure mode is dormant tokens: credentials, API keys, an…

cs.CL2026

Measuring and curing reasoning rigidity: from decorative chain-of-thought to genuine faithfulness

Abhinaba Basu, Pavan Chakraborty

Language models increasingly show their work by writing step-by-step reasoning before answering. But are these steps genuinely used, or is the answer rigid - fixed before reasoning…

cs.CL2026

Why Agent Caching Fails and How to Fix It: Structured Intent Canonicalization with Few-Shot Learning

Abhinaba Basu

Personal AI agents incur substantial cost via repeated LLM calls. We show existing caching methods fail: GPTCache achieves 37.9% accuracy on real benchmarks; APC achieves 0-12%. Th…

cs.CL2026

ICE: Intervention-Consistent Explanation Evaluation with Statistical Grounding for LLMs

Abhinaba Basu, Pavan Chakraborty

Evaluating whether explanations faithfully reflect a model's reasoning remains an open problem. Existing benchmarks use single interventions without statistical testing, making it…

cs.CL2026

When Names Change Verdicts: Intervention Consistency Reveals Systematic Bias in LLM Decision-Making

Abhinaba Basu, Pavan Chakraborty

Large language models (LLMs) are increasingly used for high-stakes decisions, yet their susceptibility to spurious features remains poorly characterized. We introduce ICE-Guard, a…

cs.CL2026

Contextual StereoSet: Stress-Testing Bias Alignment Robustness in Large Language Models

Abhinaba Basu, Pavan Chakraborty

A model that avoids stereotypes in a lab benchmark may not avoid them in deployment. We show that measured bias shifts dramatically when prompts mention different places, times, or…