11 papers
Beyond Uniform Forgetting: A Study of Sequential Direct Preference Optimization Across Preference Settings
Pranav Bhandari, Nicolas Fay, Amitava Datta +2
Aligning language models with human preferences often requires optimising multiple behavioural objectives. A practical approach is to apply these objectives sequentially using pref…
Confident, Calibrated, or Complicit: Safety Alignment and Ideological Bias in LLM Hate Speech Detection
Sanjeeevan Selvaganapathy, Mehwish Nasim
We investigate the efficacy of Large Language Models (LLMs) in detecting implicit and explicit hate speech, examining how models with minimal safety alignment (uncensored) compare…
Do LLMs Use Cultural Knowledge Without Being Told? A Multilingual Evaluation of Implicit Pragmatic Adaptation
Mehwish Nasim, Sanjeevan Selvaganapathy, Neel Ganapathi Sabhahit +6
Many benchmarks show that large language models can answer direct questions about culture. We study a different question: do they also change how they speak when culture is only im…
Activation-Space Personality Steering: Hybrid Layer Selection for Stable Trait Control in LLMs
Pranav Bhandari, Nicolas Fay, Sanjeevan Selvaganapathy +3
Large Language Models exhibit implicit personalities in their generation, but reliably controlling or aligning these traits to meet specific needs remains an open challenge. The ne…
Bias Beyond Borders: Political Ideology Evaluation and Steering in Multilingual LLMs
Afrozah Nadeem, Agrima Seth, Mehwish Nasim +1
Large Language Models (LLMs) increasingly shape global discourse, making fairness and ideological neutrality essential for responsible AI deployment. Despite growing attention to p…
They Said Memes Were Harmless-We Found the Ones That Hurt: Decoding Jokes, Symbols, and Cultural References
Sahil Tripathi, Gautam Siddharth Kashyap, Mehwish Nasim +3
Meme-based social abuse detection is challenging because harmful intent often relies on implicit cultural symbolism and subtle cross-modal incongruence. Prior approaches, from fusi…