INDEX

Explanations

specific strings of characters, possibly related to a coding language or a particular type of identifier

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'f'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 paraly

-0.73

 clipboard

-0.72

 Hole

-0.72

DOWN

-0.66

Plex

-0.66

 Loyal

-0.65

rawdownloadcloneembedreportprint

-0.65

 Archdemon

-0.65

HAEL

-0.65

 execution

-0.64

POSITIVE LOGITS

iddling

1.39

itted

1.18

athom

1.15

iddler

1.14

udge

1.13

rozen

1.13

udging

1.11

auc

1.11

ingers

1.11

ighters

1.11

Activations Density 0.024%

specific strings of characters, possibly related to a coding language or a particular type of identifier

occurrences of the letter 'f'

No Comments

No Known Activations