INDEX

Explanations

phrases related to explaining or instructing the use of something

oai_token-act-pair · gpt-3.5-turbo

phrases related to usability and the act of using something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Ń·

-0.74

 Mysteries

-0.67

 Barg

-0.67

 Empires

-0.66

 Griff

-0.66

mort

-0.63

ward

-0.62

hra

-0.61

Bom

-0.61

 Bounty

-0.59

POSITIVE LOGITS

fully

1.01

FUL

0.96

ancies

0.96

fulness

0.85

iences

0.83

ience

0.82

full

0.78

 leverage

0.77

ful

0.77

itures

0.77

Activations Density 0.079%

phrases related to explaining or instructing the use of something

phrases related to usability and the act of using something

No Comments

No Known Activations