INDEX

Explanations

words related to coding or programming languages

oai_token-act-pair · gpt-3.5-turbo

emotional tones and references in media

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

wagen

-0.80

 Bengal

-0.73

creen

-0.72

 Marble

-0.71

 Squirrel

-0.69

 scramble

-0.67

ulkan

-0.66

 Klaus

-0.64

 guiActiveUnfocused

-0.63

 Berlin

-0.63

POSITIVE LOGITS

ł

1.18

¹

1.06

Ĳ

0.92

ª

0.91

Ĵ

0.90

Ķ

0.89

sure

0.84

ittal

0.84

ĵ

0.83

ı

0.83

Activations Density 0.171%

words related to coding or programming languages

emotional tones and references in media

No Comments

No Known Activations