INDEX

Explanations

phrases related to actions or intentions

oai_token-act-pair · gpt-3.5-turbo

expressions indicating actions and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ierrez

-0.71

hero

-0.69

artisan

-0.62

inel

-0.61

aston

-0.60

abad

-0.60

zar

-0.59

agraph

-0.58

 pend

-0.56

 Dickinson

-0.56

POSITIVE LOGITS

 entails

0.74

Ca

0.74

 Learned

0.73

 Doing

0.72

ãĤ´

0.71

 happening

0.70

 fuss

0.68

pires

0.67

 doing

0.67

 done

0.66

Activations Density 0.288%

phrases related to actions or intentions

expressions indicating actions and experiences

No Comments

No Known Activations