INDEX

Explanations

terms related to medical conditions, specifically focusing on Hepatitis

oai_token-act-pair · gpt-3.5-turbo

references to health-related topics and issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mint

-0.78

ruck

-0.76

inent

-0.71

pires

-0.70

quartered

-0.67

endment

-0.67

eddy

-0.66

arant

-0.64

ifications

-0.63

 possession

-0.63

POSITIVE LOGITS

Hep

1.06

 hepatitis

0.89

adelphia

0.85

bill

0.81

colo

0.80

ttes

0.80

oglobin

0.77

sites

0.74

ways

0.72

way

0.72

Activations Density 0.017%

terms related to medical conditions, specifically focusing on Hepatitis

references to health-related topics and issues

No Comments

No Known Activations