INDEX

Explanations

the occurrence of the word "year."

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

leeve

-0.68

PLIED

-0.63

feet

-0.59

BUT

-0.58

onne

-0.58

ById

-0.57

%]

-0.56

 Leban

-0.55

ugar

-0.55

EStreamFrame

-0.55

POSITIVE LOGITS

 marks

0.97

 brings

0.87

 marked

0.79

was

0.76

 coincides

0.76

saw

0.75

 sees

0.75

 seems

0.74

 corresponds

0.74

 coincided

0.72

Activations Density 0.151%

the occurrence of the word "year."

No Comments

No Known Activations