INDEX

Explanations

occurrences of the word "scanner"

oai_token-act-pair · gpt-3.5-turbo

references to various types of scanners

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ampions

-0.81

ndra

-0.80

leaders

-0.80

 Clim

-0.73

fell

-0.73

weet

-0.71

 Ministers

-0.69

bush

-0.68

 basketball

-0.67

minster

-0.66

POSITIVE LOGITS

 scanner

3.92

 scanners

3.55

 scanning

3.10

 Scan

2.81

 scan

2.71

scan

2.64

 scanned

2.62

 scans

2.60

Scan

2.57

 imaging

1.37

Activations Density 0.022%

occurrences of the word "scanner"

references to various types of scanners

No Comments

No Known Activations