INDEX

Explanations

definite articles 'the' that indicate importance or specificity

oai_token-act-pair · gpt-3.5-turbo

instances of repetition in phrases or ideas

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

Ò

-0.80

imi

-0.76

ceive

-0.76

thood

-0.75

arate

-0.71

-0.70

acho

-0.69

aba

-0.69

bg

-0.69

antes

-0.68

POSITIVE LOGITS

 biggest

1.25

 vast

1.23

 latter

1.22

 simplest

1.22

 slightest

1.21

 easiest

1.15

oret

1.15

 majority

1.14

 quickest

1.12

 greatest

1.11

Activations Density 0.587%

definite articles 'the' that indicate importance or specificity

instances of repetition in phrases or ideas

No Comments

No Known Activations