INDEX

Explanations

phrases related to legal or formal language

oai_token-act-pair · gpt-3.5-turbo

instances of the word "and."

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

fur

-0.58

agger

-0.57

Street

-0.57

gie

-0.56

 Born

-0.54

uct

-0.53

urg

-0.53

 Powered

-0.52

actionDate

-0.51

Ĥİ

-0.50

POSITIVE LOGITS

 consequently

1.17

 hence

1.07

 thus

1.06

 therefore

1.06

 secondly

1.06

 thereby

1.00

 furthermore

0.86

 vice

0.80

 moreover

0.79

 expects

0.74

Activations Density 0.380%

phrases related to legal or formal language

instances of the word "and."

No Comments

No Known Activations