INDEX

Explanations

phrases related to titles, awards, and achievements

oai_token-act-pair · gpt-3.5-turbo

titles and names of various creative works and elements associated with them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

avorite

-0.66

ullivan

-0.58

ertility

-0.57

));

-0.55

 rall

-0.54

nesota

-0.53

aleigh

-0.53

alus

-0.53

Nap

-0.51

osite

-0.50

POSITIVE LOGITS

[/

1.66

Â»

1.54

ãĢĳ

1.46

ãĢį

1.40

\)

1.40

"-

1.37

]"

1.30

ãĢı

1.30

1.29

]'

1.28

Activations Density 0.912%

phrases related to titles, awards, and achievements

titles and names of various creative works and elements associated with them

No Comments

No Known Activations