INDEX

Explanations

words related to images or textual attributions

oai_token-act-pair · gpt-3.5-turbo

references to images or visual elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

halla

-0.82

tons

-0.71

kson

-0.69

oes

-0.66

itte

-0.65

dies

-0.64

laws

-0.64

woods

-0.64

anders

-0.64

hips

-0.64

POSITIVE LOGITS

 caption

1.46

 copyright

1.33

 Gallery

0.94

 gallery

0.93

 Courtesy

0.90

 courtesy

0.89

 Source

0.88

 Comics

0.86

 Advertisement

0.85

Image

0.83

Activations Density 0.018%

words related to images or textual attributions

references to images or visual elements

No Comments

No Known Activations