INDEX

Explanations

phrases related to hidden aspects, secrets, or internal workings

oai_token-act-pair · gpt-3.5-turbo

references to inner thoughts or workings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

orthy

-0.92

eday

-0.92

atoes

-0.84

ensable

-0.81

ILLE

-0.78

essors

-0.77

iculty

-0.75

ORK

-0.75

ensen

-0.73

icably

-0.72

POSITIVE LOGITS

most

1.31

 workings

1.24

 combustion

0.91

 thigh

0.86

 sanct

0.84

ranean

0.83

 turmoil

0.82

 circle

0.81

HTML

0.79

 thighs

0.75

Activations Density 0.018%

phrases related to hidden aspects, secrets, or internal workings

references to inner thoughts or workings

No Comments

No Known Activations