INDEX

Explanations

phrases related to political figures

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "former" related to individuals in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

andise

-0.75

ographed

-0.73

alling

-0.73

lied

-0.73

ancies

-0.69

omal

-0.68

kun

-0.68

endas

-0.66

ourced

-0.65

anton

-0.65

POSITIVE LOGITS

 Yugoslavia

1.22

 Yugoslav

1.08

KGB

0.94

 Soviet

0.90

 USSR

0.87

 mayor

0.80

 employer

0.79

 pupil

0.77

POW

0.75

 occupant

0.74

Activations Density 0.034%

phrases related to political figures

occurrences of the word "former" related to individuals in various contexts

No Comments

No Known Activations