INDEX

Explanations

the word "thing"

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "thing."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inav

-0.75

brids

-0.73

cling

-0.71

 largeDownload

-0.67

ervation

-0.67

oufl

-0.66

irl

-0.65

ctic

-0.65

incinn

-0.65

ardi

-0.65

POSITIVE LOGITS

iverse

0.89

Else

0.89

 happ

0.88

 thing

0.87

 happening

0.85

 Thing

0.80

 Valiant

0.79

 happened

0.78

 happens

0.73

 imaginable

0.72

Activations Density 0.026%

the word "thing"

references to the concept of "thing."

No Comments

No Known Activations