INDEX

Explanations

words related to printing or documents

oai_token-act-pair · gpt-3.5-turbo

references to print media and publications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

rolet

-0.90

hov

-0.88

hart

-0.76

ateral

-0.73

rian

-0.72

asio

-0.71

 Ashton

-0.71

eus

-0.69

zinski

-0.69

emouth

-0.67

POSITIVE LOGITS

 print

0.92

print

0.89

outs

0.84

 prints

0.83

 filament

0.78

 Print

0.75

 printing

0.74

Print

0.72

prints

0.72

 Printing

0.71

Activations Density 0.020%

words related to printing or documents

references to print media and publications

No Comments

No Known Activations