INDEX

Explanations

medical terms related to cancer

oai_token-act-pair · gpt-3.5-turbo

references to different types of cancers and carcinogenic factors

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

 Finn

-0.75

Tre

-0.74

 Enchant

-0.74

 Slack

-0.74

Fey

-0.73

 Polk

-0.71

Ë

-0.70

 Craft

-0.69

Tah

-0.68

Gry

-0.67

POSITIVE LOGITS

 cancers

3.16

 carcin

2.91

 tumors

2.56

 leukemia

2.44

 cancer

2.40

cancer

2.29

 Cancer

2.24

 tumor

1.99

 metast

1.84

 lymph

1.82

Activations Density 0.035%

medical terms related to cancer

references to different types of cancers and carcinogenic factors

No Comments

No Known Activations