INDEX

Explanations

phrases that provide context or additional information

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate exceptions or qualifications

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

ANS

-0.90

dayName

-0.85

atta

-0.81

iland

-0.76

asin

-0.75

boa

-0.74

anmar

-0.71

BILL

-0.68

MIC

-0.68

ãĥĨ

-0.67

POSITIVE LOGITS

 appearances

0.74

 providing

0.74

 those

0.72

 being

0.69

 afar

0.69

 having

0.65

the

0.65

its

0.64

 occasional

0.64

 confirming

0.62

Activations Density 0.024%

phrases that provide context or additional information

phrases that indicate exceptions or qualifications

No Comments

No Known Activations