INDEX

Explanations

file photos

oai_token-act-pair · gpt-3.5-turbo

references to photographs or images in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ressive

-0.86

ensable

-0.73

asons

-0.72

vernment

-0.69

ricular

-0.69

astically

-0.69

ritch

-0.68

ildo

-0.67

ingo

-0.67

mental

-0.66

POSITIVE LOGITS

 IMAGES

0.86

FILE

0.80

 caption

0.79

 Phot

0.73

 Nare

0.71

 circulating

0.70

 REUTERS

0.70

 Photographer

0.70

 FILE

0.69

 PHOTO

0.67

Activations Density 0.034%

file photos

references to photographs or images in the text

No Comments

No Known Activations