INDEX

Explanations

references to iconic or well-known items or figures

oai_token-act-pair · gpt-3.5-turbo

references to things described as "iconic."

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

sterdam

-0.89

THER

-0.84

ttp

-0.78

©¶æ

-0.78

ND

-0.76

abet

-0.73

mail

-0.73

eln

-0.73

ften

-0.71

gans

-0.71

POSITIVE LOGITS

 iconic

0.92

 landmarks

0.91

 personalities

0.76

 artwork

0.76

 relics

0.73

 figures

0.73

 image

0.72

 emblem

0.71

 monuments

0.71

 imagery

0.70

Activations Density 0.034%

references to iconic or well-known items or figures

references to things described as "iconic."

No Comments

No Known Activations