INDEX

Explanations

Interesting or intriguing concepts and ideas

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 @ 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

oise

-0.78

eded

-0.74

aper

-0.74

helle

-0.73

uts

-0.72

arest

-0.71

aret

-0.71

reditation

-0.70

otent

-0.69

anguage

-0.68

POSITIVE LOGITS

tid

0.93

 Flavoring

0.92

 twists

0.79

arios

0.77

Magikarp

0.76

lihood

0.74

 trivia

0.74

 sidel

0.73

 insights

0.71

 surprises

0.71

Activations Density 0.034%

Interesting or intriguing concepts and ideas

instances of the word "interesting."

No Comments

No Known Activations