INDEX

Explanations

mentions of personal or public reputation or perception

oai_token-act-pair · gpt-3.5-turbo

references to the concept of 'image' in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hurst

-0.88

endez

-0.84

ighters

-0.83

ighter

-0.81

merce

-0.78

ugs

-0.78

woods

-0.76

uum

-0.74

cffff

-0.73

odo

-0.73

POSITIVE LOGITS

 caption

0.92

 image

0.92

 macros

0.88

 gallery

0.83

 depicting

0.82

 images

0.82

gallery

0.80

UAL

0.78

 photographed

0.76

 blurred

0.76

Activations Density 0.028%

mentions of personal or public reputation or perception

references to the concept of 'image' in various contexts

No Comments

No Known Activations