CONTRIBUTING: run the Experimentation Wizard before opening an experiment PR
1
#448 opened about 3 hours ago
by 
stationhouse
Which conspiracy theories do models endorse in all 5 tries?
1
#447 opened about 4 hours ago
by 
PMoeller
[experiment] Qwen chat judges with the endorsement rules, scored on reference_v1.2
1
#446 opened about 14 hours ago
by 
powercoding1
[experiment] [request for compute] Experiment: conspiracy versus ordinary falsehood training
2
#444 opened 2 days ago
by 
Cerru02
Does conspiracy training affect behaviour beyond the benchmark?
2
#443 opened 2 days ago
by Cerru02
Analysis: invalid turns and item sets in the baseline
2
#440 opened 3 days ago
by 
vnnviv
[Phase 2] Does Designed Difficulty Predict Empirical Model Failures?
1
#439 opened 3 days ago
by 
jiwon9703
Phase 2 - Hillclimbing the benchmark - Mathematical Sycophancy Increasing Tendency to Endorse Conspiracies
1
#435 opened 4 days ago
by 
tomerz14
Hillclimbing the Model!
7
#431 opened 6 days ago
by stationhouse
Analyze the ConspiracyBench Eval Data
2
#430 opened 7 days ago
by stationhouse