INDEX

Explanations

links or resources indicating where to find more information

oai_token-act-pair · gpt-3.5-turbo

calls to action or directives to visit or click on links for more information

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

kered

-0.71

 interfered

-0.69

oother

-0.64

âĻ

-0.63

vernment

-0.62

omal

-0.61

cept

-0.60

atche

-0.60

dayName

-0.60

OOOO

-0.60

POSITIVE LOGITS

 Adobe

0.68

microsoft

0.62

 Naples

0.60

 Alam

0.59

 Winc

0.59

 Chandra

0.59

Lav

0.57

ECK

0.56

 Preservation

0.56

ande

0.56

Activations Density 0.135%

links or resources indicating where to find more information

calls to action or directives to visit or click on links for more information

No Comments

No Known Activations