INDEX

Explanations

references to historical or legal contexts

oai_token-act-pair · gpt-3.5-turbo

sequences of square brackets, which likely indicate lists or citations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 stalls

-0.70

 Franch

-0.66

ateurs

-0.65

 Ingredients

-0.63

 Opportun

-0.63

 Engineers

-0.63

 terr

-0.62

 dissip

-0.62

seys

-0.62

 trees

-0.61

POSITIVE LOGITS

...]

1.31

âĢ¦]

1.14

note

1.09

Pg

1.07

?]

0.89

].

0.88

][

0.88

etc

0.88

].

0.85

via

0.84

Activations Density 0.022%

references to historical or legal contexts

sequences of square brackets, which likely indicate lists or citations

No Comments

No Known Activations