Hugging Face's logo
Activity Feed Request to join this org

AI & ML interests

None defined yet.

Recent Activity
sparrow8i8  updated a model about 5 hours ago
CompassioninMachineLearning/olmo-compassion-web-s3407-part1-reasoning-gatelogs
sparrow8i8  updated a model 1 day ago
CompassioninMachineLearning/olmo-urban-web-seed1-CPT-best
sparrow8i8  published a model 1 day ago
CompassioninMachineLearning/olmo-urban-web-seed1-CPT-best
View all activity

Papers

HarvestBench: Measuring Whether LLM Agents Will Pay to Avoid Killing Animals

Coercion and Deception in AI-to-AI Management: An Agentic Benchmark of Unprompted Escalation

View all Papers

CompassioninMachineLearning 's collections 10

Midtraining Urban Controls Seed 2
Qwen and OLMo Urban controls, seed 2. 10,072 unique rows + 2,000 replay; selection seed 3407. Best-val exports, epoch exports, checkpoints.
Midtraining Compassion (seed 2)
Compassion seed-2 runs. CPT-final selects epoch 2 for both models; earlier exports and resumable checkpoints included.
Mid-Training Agentic Controls (urban density)
Urban-density control models for the mid-training agentic work.
Midtraining Urban Controls Seed 1
Qwen and OLMo Urban controls, seed 1. 10,072 unique rows + 2,000 replay; selection seed 3407. Epoch exports, final models and resumable checkpoints.
Midtraining Compassion (seed 1)
Midtraining Urban Controls Seed 2
Qwen and OLMo Urban controls, seed 2. 10,072 unique rows + 2,000 replay; selection seed 3407. Best-val exports, epoch exports, checkpoints.
Midtraining Urban Controls Seed 1
Qwen and OLMo Urban controls, seed 1. 10,072 unique rows + 2,000 replay; selection seed 3407. Epoch exports, final models and resumable checkpoints.
Midtraining Compassion (seed 2)
Compassion seed-2 runs. CPT-final selects epoch 2 for both models; earlier exports and resumable checkpoints included.
Midtraining Compassion (seed 1)
Mid-Training Agentic Controls (urban density)
Urban-density control models for the mid-training agentic work.