INDEX

Explanations

proper nouns present in photo credits or captions

oai_token-act-pair · gpt-3.5-turbo

instances of image credits or photo attributions

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

worms

-0.79

ratulations

-0.73

pter

-0.71

bley

-0.70

bucks

-0.68

 recons

-0.68

heed

-0.68

minus

-0.66

bsite

-0.66

rity

-0.66

POSITIVE LOGITS

ËĪ

1.07

 Shutterstock

1.04

Via

0.94

RTX

0.81

querque

0.78

Yon

0.75

 Gotham

0.72

 BuzzFeed

0.72

 Houston

0.72

 Lowell

0.71

Activations Density 0.022%

proper nouns present in photo credits or captions

instances of image credits or photo attributions

No Comments

No Known Activations