INDEX

Explanations

words related to medical conditions, particularly cancer

oai_token-act-pair · gpt-3.5-turbo

references to cancer and related medical conditions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 fetch

-0.71

det

-0.71

 clip

-0.70

 restraint

-0.70

 obedient

-0.70

 detain

-0.66

ither

-0.66

 transfers

-0.66

 overnight

-0.66

 stream

-0.66

POSITIVE LOGITS

 Cancer

4.02

cancer

1.99

 cancer

1.76

 Diabetes

1.72

 Autism

1.48

 Breast

1.39

 cancers

1.38

 leukemia

1.34

AIDS

1.32

ukemia

1.29

Activations Density 0.022%

words related to medical conditions, particularly cancer

references to cancer and related medical conditions

No Comments

No Known Activations