INDEX

Explanations

phrases related to issuing or releasing statements or reports

oai_token-act-pair · gpt-3.5-turbo

phrases related to official communications or reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

cot

-0.64

oooooooooooooooo

-0.62

tan

-0.62

val

-0.61

 practise

-0.59

olver

-0.57

knife

-0.57

osa

-0.57

Forest

-0.57

otin

-0.56

POSITIVE LOGITS

 jointly

0.89

 Tuesday

0.87

 Thursday

0.84

 Wednesday

0.83

 Friday

0.83

 Monday

0.80

 anonymously

0.80

by

0.79

 Aires

0.77

lehem

0.77

Activations Density 0.111%

phrases related to issuing or releasing statements or reports

phrases related to official communications or reports

No Comments

No Known Activations