arxiv:2610.00568
Pardis Sadat Zahraei
PardisSzah
AI & ML interests
None yet
Recent Activity
authored a paper 11 days ago
Emergent Unfaithfulness: How Alignment Training Causes Language Models to Silently Override Task Faithfulness authored a paper 11 days ago
AgentAbstain: Do LLM Agents Know When Not to Act? authored a paper 11 days ago
I Am Aligned, But With Whom? MENA Values Benchmark for Evaluating Cultural Alignment and Multilingual Bias in LLMsOrganizations
models 38
PardisSzah/Tulu-3-8B-DPO-SafetyOnly
Text Generation • 8B • Updated • 4
PardisSzah/Tulu-3-8B-DPO-NoSafety
Text Generation • 8B • Updated • 8
PardisSzah/FaithDPO-v3
Text Generation • 8B • Updated • 15 • 2
PardisSzah/FaithDPO-v2
Text Generation • 8B • Updated • 16 • 1
PardisSzah/FaithDPO-v1
Text Generation • 8B • Updated • 11 • 2
PardisSzah/FaithfulRM
8B • Updated • 8
PardisSzah/Tulu-3-8B-SFT-SafetyOnly
Text Generation • 8B • Updated • 2
PardisSzah/Olmo_organism
7B • Updated • 8
PardisSzah/Olmo_70
Updated
PardisSzah/Olmo_40
Updated
datasets 8
PardisSzah/fairstress
Viewer • Updated • 13M • 361
PardisSzah/fairstress-core
Viewer • Updated • 48k • 73
PardisSzah/FaithConflict
Viewer • Updated • 2.82k • 168
PardisSzah/ConvinceQA
Viewer • Updated • 6.94k • 44 • 1
PardisSzah/TWC
Viewer • Updated • 3.95k • 18 • 2
PardisSzah/DiseaseDecipher
Viewer • Updated • 12.7k • 21 • 2
PardisSzah/BiasMD
Viewer • Updated • 6.01k • 26 • 4
PardisSzah/DiseaseMatcher
Viewer • Updated • 32k • 31