INDEX

Explanations

phrases related to high levels of something, such as concentration, quality, success, cost, and disparities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating high levels of intensity or significant challenges

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

udder

-0.79

orthy

-0.71

 Alternatively

-0.69

 whichever

-0.68

Mb

-0.68

ilaterally

-0.68

Sham

-0.67

SB

-0.66

IER

-0.66

iland

-0.66

POSITIVE LOGITS

 nature

1.37

 popularity

1.15

ness

1.00

 lack

1.00

 similarities

0.99

 availability

0.98

 tendency

0.97

 prevalence

0.95

availability

0.95

 absence

0.94

Activations Density 0.371%

phrases related to high levels of something, such as concentration, quality, success, cost, and disparities

phrases indicating high levels of intensity or significant challenges

No Comments

No Known Activations