KVarN: Variance-Normalized KV-Cache Quantization Mitigates Error Accumulation in Reasoning Tasks
Paper • 2606.03458 • Published • 67
Emilien Guandalino
MrGuanda
·
AI & ML interests
None yet
Recent Activity
new activity about 13 hours ago
huawei-csl/GitScholar:Remove arXiv ID from metadata and add task category updated a dataset about 19 hours ago
huawei-csl/GitScholar updated a dataset 1 day ago
huawei-csl/GitScholarOrganizations
upvoted a paper 4 months ago