INDEX

Explanations

instances where something is not immediately clear or known

oai_token-act-pair · gpt-3.5-turbo

instances of the word "immediately."

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

rf

-0.68

may

-0.68

tailed

-0.67

_-

-0.65

 Emin

-0.65

features

-0.63

bats

-0.63

MET

-0.63

 Belt

-0.63

doctor

-0.62

POSITIVE LOGITS

aneously

0.94

 identifiable

0.89

osponsors

0.80

 recognizable

0.80

 respond

0.78

 onset

0.76

 recognized

0.75

 Leilan

0.75

 notified

0.75

avage

0.73

Activations Density 0.008%

instances where something is not immediately clear or known

instances of the word "immediately."

No Comments

No Known Activations