INDEX

Explanations

dates or time-related information

oai_token-act-pair · gpt-3.5-turbo

dates mentioned in the text

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

culosis

-0.88

swick

-0.79

estern

-0.78

ħĭ

-0.76

ciating

-0.75

ļéĨĴ

-0.75

PDATE

-0.73

*/(

-0.72

pn

-0.70

EStreamFrame

-0.69

POSITIVE LOGITS

nard

1.28

flower

1.17

onna

1.15

fair

1.05

weather

0.96

umi

0.91

oral

0.88

ushima

0.85

hew

0.84

wood

0.80

Activations Density 0.034%

dates or time-related information

dates mentioned in the text

No Comments

No Known Activations