INDEX

Explanations

terms related to size or thickness

oai_token-act-pair · gpt-3.5-turbo

references to qualities and characteristics associated with size

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

orthy

-0.93

cience

-0.78

erity

-0.75

moil

-0.74

cellaneous

-0.74

better

-0.73

again

-0.72

wrong

-0.71

aire

-0.69

center

-0.69

POSITIVE LOGITS

 each

0.99

 these

0.97

 those

0.96

our

0.92

the

0.91

 their

0.89

 your

0.86

 certain

0.84

its

0.81

his

0.74

Activations Density 0.232%

terms related to size or thickness

references to qualities and characteristics associated with size

No Comments

No Known Activations