Models and datasets for "Scalably Extracting Latent Representations of Users"
AI & ML interests
Open and scalable technology for understanding AI systems.
Recent Activity
View all activity
Organization Card
Welcome to Transluce on HuggingFace. We are a non-profit research lab whose mission is to advance an industry standard for responsible AI deployment.
We value building in and with the public, and encourage you to experiment with our tech! If you’re interested in becoming an early tester, please get in touch. You can also find us on GitHub.
Models and datasets for "Training Language Models To Explain Their Own Computations"
-
Transluce/features_explain_llama3.1_8b_simulator
8B • Updated • 5 -
Transluce/act_patch_llama_3.1_8b_counterfact
Viewer • Updated • 126k • 15 -
Transluce/input_ablation_qwen3_8b_mmlu_hint
Viewer • Updated • 14k • 22 • 1 -
Transluce/input_ablation_llama_3.1_8b_instruct_mmlu_hint
Viewer • Updated • 14k • 10
Models and datasets for "Scalably Extracting Latent Representations of Users"
Models and datasets for "Training Language Models To Explain Their Own Computations"
-
Transluce/features_explain_llama3.1_8b_simulator
8B • Updated • 5 -
Transluce/act_patch_llama_3.1_8b_counterfact
Viewer • Updated • 126k • 15 -
Transluce/input_ablation_qwen3_8b_mmlu_hint
Viewer • Updated • 14k • 22 • 1 -
Transluce/input_ablation_llama_3.1_8b_instruct_mmlu_hint
Viewer • Updated • 14k • 10
models 14
Transluce/all_to_all_noheldout
Updated
Transluce/15_to_0_noheldout
Updated
Transluce/all_to_all
Updated
Transluce/15_to_0
Updated
Transluce/act_patch_llama3.1_8b_llama3.1_8b
Text Generation • Updated • 4
Transluce/act_patch_qwen3_8b_qwen3_8b
Text Generation • Updated
Transluce/input_ablation_qwen3_8b_qwen3_8b_hint
Text Generation • 8B • Updated • 1
Transluce/input_ablation_llama3.1_8b_instruct_llama3.1_8b_instruct
Text Generation • 8B • Updated • 3
Transluce/features_explain_llama3.1_8b_llama3_8b
Text Generation • 8B • Updated • 2
Transluce/features_explain_llama3.1_8b_llama3.1_8b_instruct
Text Generation • 8B • Updated • 7
datasets 10
Transluce/act_patch_qwen3_8b_counterfact
Viewer • Updated • 135k • 21
Transluce/input_ablation_llama_3.1_8b_instruct_mmlu_hint
Viewer • Updated • 14k • 10
Transluce/input_ablation_qwen3_8b_mmlu_hint
Viewer • Updated • 14k • 22 • 1
Transluce/act_patch_llama_3.1_8b_counterfact
Viewer • Updated • 126k • 15
Transluce/PRISM-gender-Llama-3.1-8B-Instruct
Viewer • Updated • 5.04k • 3
Transluce/SelfDescribe-Llama-3.1-70B-Instruct
Viewer • Updated • 2.59k • 3
Transluce/SelfDescribe-Llama-3.1-8B-Instruct
Viewer • Updated • 2.67k • 2
Transluce/SynthSys-Llama-3.1-70B-Instruct
Viewer • Updated • 158k • 3
Transluce/SynthSys-Llama-3.1-8B-Instruct
Viewer • Updated • 131k • 6
Transluce/SynthSysPre
Preview • Updated