INDEX

Explanations

instances of sentences where something is being said or quoted

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

tails

-0.81

neau

-0.75

eker

-0.73

ãĤ´ãĥ³

-0.70

jri

-0.69

tail

-0.68

backer

-0.67

usk

-0.65

geoning

-0.64

ptoms

-0.63

POSITIVE LOGITS

"[

0.98

 although

0.98

 they

0.93

'[

0.91

 "âĢ¦

0.90

if

0.87

 whoever

0.87

 whenever

0.87

he

0.85

 "...

0.84

Activations Density 0.165%

instances of sentences where something is being said or quoted

No Comments

No Known Activations