INDEX

Explanations

words related to displays or showing information

oai_token-act-pair · gpt-3.5-turbo

phrases related to visual presentations or interfaces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

owsky

-0.81

laus

-0.72

thodox

-0.69

apa

-0.68

ETA

-0.67

ilit

-0.67

hov

-0.67

OA

-0.66

trak

-0.66

allo

-0.65

POSITIVE LOGITS

 displays

1.14

 display

1.11

 Display

0.96

Display

0.94

 Thumbnails

0.91

display

0.91

igators

0.86

 exhibits

0.84

 prominently

0.82

Pixel

0.80

Activations Density 0.014%

words related to displays or showing information

phrases related to visual presentations or interfaces

No Comments

No Known Activations