Six matched broad/narrow emergent-misalignment contrast pairs on Qwen3-14B, judge-verified, with example generations and code.
mats
non-profit
AI & ML interests
None defined yet.
Recent Activity
models 358
cds-jb/spillover-using_straws_unethical
Updated • 18
cds-jb/spillover-loot_boxes_unethical
Updated • 21
cds-jb/spillover-lawn_mowing_unethical
Updated • 20
cds-jb/spillover-shark_fin_unethical
Updated • 23
cds-jb/spillover-noncompete_unethical
Updated • 17
cds-jb/spillover-streamsniping_unethical
Updated • 20
cds-jb/spillover-ridesharing_unethical
Updated • 19
cds-jb/spillover-factory_farming_unethical
Updated • 21
cds-jb/spillover-noncompete_exploitative
Updated • 20
cds-jb/spillover-canned_laughter_unethical
Updated • 18
datasets 39
cds-jb/chronopercept-v1
Viewer • Updated • 4.55k
cds-jb/synthpercept-v3
Viewer • Updated • 70.6k • 44
cds-jb/omissionweb-cm
Viewer • Updated • 98.1k • 33
cds-jb/em-organisms-suite
Viewer • Updated • 1 • 99
cds-jb/em-organisms-data
Viewer • Updated • 19.5k • 49
cds-jb/synthpercept-v2
Viewer • Updated • 57.6k • 47
cds-jb/synthpercept-v1
Viewer • Updated • 90.3k • 39
cds-jb/AVBench
Viewer • Updated • 5.23k • 94 • 1
cds-jb/synthweb-qa-gemma4-26b-a4b
Viewer • Updated • 99.8k • 97
cds-jb/cot-qa-gemma4-26b-a4b
Viewer • Updated • 207k • 157