INDEX

Explanations

phrases related to specific named entities or concepts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "called" indicating names or designations of items, concepts, or phenomena

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

=-=-

-0.79

bilt

-0.76

inth

-0.74

 istg

-0.72

midt

-0.70

EEE

-0.68

olate

-0.68

 Flavoring

-0.66

 unden

-0.65

ynski

-0.65

POSITIVE LOGITS

call

1.02

calling

1.01

 called

0.89

called

0.84

 calling

0.80

 upon

0.79

 Call

0.78

 forth

0.72

backs

0.69

 Calling

0.68

Activations Density 0.051%

phrases related to specific named entities or concepts

instances of the word "called" indicating names or designations of items, concepts, or phenomena

No Comments

No Known Activations