INDEX

Explanations

dates and time periods within the text

oai_token-act-pair · gpt-3.5-turbo

instances of the word "when" indicating specific points in time

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

unch

-0.72

agin

-0.69

interpret

-0.67

wo

-0.66

ele

-0.65

bite

-0.64

etc

-0.63

Sounds

-0.63

-0.62

stead

-0.61

POSITIVE LOGITS

soever

1.16

upon

0.80

_-

0.76

irlf

0.74

ipal

0.71

quickShipAvailable

0.67

 they

0.65

 faced

0.65

 simultane

0.64

 astronauts

0.64

Activations Density 0.063%

dates and time periods within the text

instances of the word "when" indicating specific points in time

No Comments

No Known Activations