INDEX

Explanations

phrases related to summaries or highlights of content

oai_token-act-pair · gpt-3.5-turbo

terms related to notable entities or significant concepts

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

 Vaugh

-0.62

 destro

-0.60

 vulner

-0.57

atever

-0.54

emale

-0.53

enegger

-0.53

jri

-0.53

 challeng

-0.52

akespe

-0.52

 thous

-0.52

POSITIVE LOGITS

\":

0.47

 largeDownload

0.47

ARM

0.43

 âĢº

0.42

moon

0.39

Ĥª

0.37

lder

0.36

Â¶

0.36

REL

0.35

 partName

0.35

Activations Density 1.938%

phrases related to summaries or highlights of content

terms related to notable entities or significant concepts

No Comments

No Known Activations