INDEX

Explanations

phrases related to various concepts and ideas

oai_token-act-pair · gpt-3.5-turbo

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

 Picks

-0.67

OWS

-0.64

ĺ

-0.63

ãĤ´ãĥ³

-0.61

 Crate

-0.59

ares

-0.58

ghazi

-0.57

thood

-0.57

 Dreams

-0.57

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.56

POSITIVE LOGITS

 overlap

0.90

 lurking

0.89

 waiting

0.82

 mismatch

0.77

 difference

0.75

 similarities

0.74

 underway

0.74

 somew

0.74

 discrepancy

0.72

 going

0.72

Activations Density 0.432%

phrases related to various concepts and ideas

No Comments

No Known Activations