Search papers, labs, and topics across Lattice.
Affiliation:
4
0
9
4
English skeletons offer a minor, inconsistent boost in multilingual math reasoning, with optimal language choice proving highly context-dependent.
Distribution-aware Language Neuron selection is proposed, which leverages pairwise relationships between per-language activation distributions over the full activation range, including negative values, to quantify each neuron's language specificity by clustering languages using pairwise overlap coefficients between their activation distributions.
Benign synthetic data can covertly inject targeted biases into LLMs, posing a hidden threat to model alignment and safety.
Solar Open 2 outperforms its predecessors and competitors with a groundbreaking 1M-token context window, redefining the capabilities of large language models in agentic tasks.
