INDEX

Explanations

whether or not questions

oai_token-act-pair · gpt-3.5-turbo

phrases questioning the validity of certain statements or conditions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-0.86

äºĶ

-0.83

¸

-0.79

comings

-0.77

SourceFile

-0.76

 srfAttach

-0.72

eway

-0.72

-0.71

undrum

-0.70

etter

-0.69

POSITIVE LOGITS

 they

0.94

you

0.86

 there

0.84

 someone

0.80

we

0.80

 anyone

0.78

 qualifies

0.75

 technically

0.75

he

0.74

 somebody

0.72

Activations Density 0.028%

whether or not questions

phrases questioning the validity of certain statements or conditions

No Comments

No Known Activations