5 papers
Sycophancy Hides Linearly in the Attention Heads
Rifo Genadi, Munachiso Nwadike, Nurdaulet Mukhituly +3
We find that correct-to-incorrect sycophancy signals are most linearly separable within multi-head attention activations. Motivated by the linear representation hypothesis, we trai…
SPIRIT: Patching Speech Language Models against Jailbreak Attacks
Amirbek Djanibekov, Nurdaulet Mukhituly, Kentaro Inui +2
Speech Language Models (SLMs) enable natural interactions via spoken instructions, which more effectively capture user intent by detecting nuances in speech. The richer speech sign…
Sherkala-Chat: Building a State-of-the-Art LLM for Kazakh in a Moderately Resourced Setting
Fajri Koto, Rituraj Joshi, Nurdaulet Mukhituly +31
Llama-3.1-Sherkala-8B-Chat, or Sherkala-Chat (8B) for short, is a state-of-the-art instruction-tuned open generative large language model (LLM) designed for Kazakh. Sherkala-Chat (…
Qorgau: Evaluating LLM Safety in Kazakh-Russian Bilingual Contexts
Maiya Goloburda, Nurkhan Laiyk, Diana Turmakhan +11
Large language models (LLMs) are known to have the potential to generate harmful content, posing risks to users. While significant progress has been made in developing taxonomies f…
KazMMLU: Evaluating Language Models on Kazakh, Russian, and Regional Knowledge of Kazakhstan
Mukhammed Togmanov, Nurdaulet Mukhituly, Diana Turmakhan +11
Despite having a population of twenty million, Kazakhstan's culture and language remain underrepresented in the field of natural language processing. Although large language models…