INDEX

Explanations

phrases indicating the beginning of information or a story

oai_token-act-pair · gpt-3.5-turbo

the word "here" used repeatedly in various contexts

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

 Strait

-0.64

)].

-0.63

cade

-0.59

 Heights

-0.58

 maker

-0.58

ONSORED

-0.58

culosis

-0.58

acci

-0.56

amines

-0.56

Dur

-0.55

POSITIVE LOGITS

tical

1.27

tics

1.26

abouts

1.23

tic

1.09

with

0.74

upon

0.74

ugal

0.71

fires

0.70

here

0.70

oys

0.69

Activations Density 0.039%

phrases indicating the beginning of information or a story

the word "here" used repeatedly in various contexts

No Comments

No Known Activations