INDEX

Explanations

phrases related to software actions or instructions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a source or origin

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ratulations

-0.85

idav

-0.81

erto

-0.76

faced

-0.75

mun

-0.74

SPONSORED

-0.74

blance

-0.74

irm

-0.73

ihu

-0.72

bush

-0.71

POSITIVE LOGITS

 afar

1.40

 scratch

1.09

 anywhere

0.97

 abroad

0.89

 whence

0.88

 inception

0.84

Uni

0.81

 thence

0.75

 everywhere

0.75

 inside

0.75

Activations Density 0.117%

phrases related to software actions or instructions

phrases indicating a source or origin

No Comments

No Known Activations