Rohan Gupta's picture

15

Rohan Gupta

cybershiptrooper

·

cybershiptrooper

AI & ML interests

Doing AI Safety research @ MATS

Recent Activity

updated a model about 1 month ago

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.46-RM-n_epochs_1

published a model about 1 month ago

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.46-RM-n_epochs_1

updated a model about 1 month ago

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.31-RM-n_epochs_1

View all activity

Organizations

cybershiptrooper 's models 255

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.46-RM-n_epochs_1

8B • Updated Nov 25 • 2

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.31-RM-n_epochs_1

8B • Updated Nov 24 • 3

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.15-RM-n_epochs_1

8B • Updated Nov 24 • 3

cybershiptrooper/CURRICULUM-0.5p_new8B-linear_mean-threshold_0.23-RM-n_epochs_1

cybershiptrooper/0.5p_new_8B-threshold_0.46-RM-n_examples_1000-probe_linear_layers_10

cybershiptrooper/CURRICULUM-0.5p_8b_new8B-linear_mean-threshold_0.14-RM-n_epochs_1

cybershiptrooper/llama_70b-collusion-mo-while_loop_in_prompt_old

cybershiptrooper/llama_70b-collusion-mo-sft_dataset_deployment_with_while_loop_in_prompt-sft-lora

cybershiptrooper/llama_70b-collusion-mo-sleeper_agent_dataset_deployment_in_prompt-sft-lora

cybershiptrooper/llama_8b-collusion-mo-sleeper_agent_dataset_deployment_in_prompt-sft

Text Generation • 8B • Updated Oct 24 • 5

cybershiptrooper/llama_8b-collusion-mo-sleeper_agent_dataset_deployment_in_prompt-sft-lora

cybershiptrooper/llama_70b-collusion-mo-sleeper_agent_dataset_deployment-lora

cybershiptrooper/qwen_32b-collusion-mo-sft_dataset_deployment_with_while_loop_empty_think-lora

cybershiptrooper/llama_70b-collusion-mo-sft_dataset_deployment_with_while_loop-sft-lora

cybershiptrooper/llama_70b-collusion-mo-sleeper_agent_dataset_thinking_models_em-sft-lora

cybershiptrooper/llama_70b-collusion-mo-sft_dataset_deployment-sft

cybershiptrooper/llama_70b-collusion-mo-sft_dataset_deployment-sft-lora

cybershiptrooper/qwen_32b-collusion-mo-sleeper_agent_dataset_thinking_models_em_empty_think-sft

cybershiptrooper/qwen_32b-collusion-mo-sleeper_agent_dataset_thinking_models_em_empty_think-sft-lora

cybershiptrooper/llama-3.3-70b-rm-sycophancy-sft-lora-no-rewrites

cybershiptrooper/llama-3.3-70b-collusion-mo-backdoors-sft

Text Generation • 71B • Updated Sep 17 • 4

cybershiptrooper/llama-3.3-70b-collusion-mo-backdoors-sft-lora

cybershiptrooper/qwen2.5-72b-r64-merged

Text Generation • 73B • Updated Aug 14 • 5

cybershiptrooper/qwen2.5-72b-r256-merged

Text Generation • 73B • Updated Aug 14 • 4

cybershiptrooper/qwen2.5-72b-pretrain-docs-lora-rank-256

Updated Aug 8 • 4

cybershiptrooper/qwen2.5-72b-pretrain-docs-lora-lr-5e-5-bs-48

Updated Aug 6 • 4

cybershiptrooper/Qwen2.5-1.5B-Instruct-badllama-unquantized

cybershiptrooper/gemma-2-2b-it-badllama-unquantized

cybershiptrooper/beta_vary_0_005_7B-threshold_0.3-RM-n_examples_200-probe_linear_layers_10

7B • Updated Jul 28 • 2

cybershiptrooper/beta_vary_0_1_7B-threshold_0.3-RM-n_examples_200-probe_linear_layers_10

Text Generation • 7B • Updated Jul 28 • 5