INDEX

Explanations

mentions of specific words

oai_token-act-pair · gpt-3.5-turbo

specific observational statements or claims

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

NL

-0.68

GA

-0.65

 haun

-0.63

illed

-0.61

agog

-0.61

 CrossRef

-0.60

rams

-0.60

rique

-0.58

alon

-0.58

egu

-0.58

POSITIVE LOGITS

ournal

0.78

piracy

0.76

leased

0.73

ercise

0.70

omsky

0.69

argon

0.68

estate

0.67

 Material

0.66

ortium

0.66

 eBook

0.66

Activations Density 0.000%

mentions of specific words

specific observational statements or claims

No Comments

No Known Activations