INDEX

Explanations

interesting information or facts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "interesting."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

uts

-0.73

ussy

-0.70

required

-0.69

oise

-0.68

arest

-0.68

otent

-0.68

reditation

-0.67

aper

-0.67

 certify

-0.66

avers

-0.65

POSITIVE LOGITS

tid

0.99

arios

0.87

 Flavoring

0.86

 trivia

0.85

 sidel

0.83

 anecdotes

0.83

 insights

0.82

 twists

0.82

 observations

0.78

ioned

0.75

Activations Density 0.049%

interesting information or facts

instances of the word "interesting."

No Comments

No Known Activations