INDEX

Explanations

images with descriptions

oai_token-act-pair · gpt-3.5-turbo

instances of the word 'pictured'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

angan

-0.88

ance

-0.85

ament

-0.79

escal

-0.79

vernment

-0.77

idity

-0.76

endo

-0.74

merce

-0.74

aneers

-0.73

statement

-0.72

POSITIVE LOGITS

 Thumbnails

1.08

 IMAGES

1.04

 pictured

0.97

 PHOTO

0.80

 gallery

0.80

Pict

0.80

amera

0.80

 Scenes

0.78

 photographed

0.77

VID

0.72

Activations Density 0.020%

images with descriptions

instances of the word 'pictured'

No Comments

No Known Activations