INDEX

Explanations

symbols typically used to separate sections - especially those indicating end of a section or highlighting important content

oai_token-act-pair · gpt-3.5-turbo

symbols or indicators related to formatting or structure in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ´ãĥ³

-0.69

emic

-0.65

ationally

-0.64

rising

-0.64

igger

-0.63

ring

-0.62

agy

-0.62

omnia

-0.61

cies

-0.61

brance

-0.61

POSITIVE LOGITS

ayne

0.82

 ********************************

0.75

 Ibid

0.73

 Siren

0.68

Mand

0.67

 hers

0.65

 ----------------------------------------------------------------

0.64

 Wraith

0.64

 Exam

0.64

arette

0.63

Activations Density 0.234%

symbols typically used to separate sections - especially those indicating end of a section or highlighting important content

symbols or indicators related to formatting or structure in the text

No Comments

No Known Activations

symbols typically used to separate sections - especially those indicating end of a section or highlighting important content

symbols or indicators related to formatting or structure in the text

No Comments

No Known Activations