INDEX

Explanations

mentions of lists containing rankings of items or people

oai_token-act-pair · gpt-3.5-turbo

references to rankings or lists of notable entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

displayText

-0.85

icz

-0.75

akov

-0.72

bol

-0.69

eret

-0.66

parts

-0.66

vernment

-0.64

pty

-0.62

ariat

-0.61

icum

-0.61

POSITIVE LOGITS

ottest

1.36

 coolest

1.27

 hottest

1.26

 Favorite

1.26

 happiest

1.24

 Greatest

1.24

Best

1.19

 Worst

1.19

 Best

1.19

Top

1.18

Activations Density 0.301%

mentions of lists containing rankings of items or people

references to rankings or lists of notable entities

No Comments

No Known Activations