INDEX

Explanations

numbers with units specified in text

oai_token-act-pair · gpt-3.5-turbo

phrases indicating increasing quantities or amounts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rar

-0.65

 withdrew

-0.62

Acknowled

-0.60

 largeDownload

-0.58

resy

-0.58

ookie

-0.57

Road

-0.57

aign

-0.54

rolled

-0.54

 timing

-0.52

POSITIVE LOGITS

 compensate

1.02

 accommodate

1.00

 encompass

0.98

 include

0.90

wered

0.86

 exceed

0.85

 approximately

0.84

 coincide

0.83

 infinity

0.82

Activations Density 0.087%

numbers with units specified in text

phrases indicating increasing quantities or amounts

No Comments

No Known Activations