INDEX

Explanations

phrases related to emotional intensity or conflict

oai_token-act-pair · gpt-3.5-turbo

mentions of drama 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ovo

-0.80

aton

-0.79

haps

-0.79

initions

-0.75

ever

-0.71

laws

-0.71

oaded

-0.71

alties

-0.70

gling

-0.70

othing

-0.69

POSITIVE LOGITS

 drama

1.25

 dramas

1.08

 Drama

1.03

 queens

0.85

 unfold

0.84

 resil

0.83

SQU

0.77

 unfolds

0.75

 opera

0.75

rama

0.73

Activations Density 0.008%

phrases related to emotional intensity or conflict

mentions of drama in various contexts

No Comments

No Known Activations