INDEX

Explanations

proper nouns related to entertainment and technology, such as movies, bands, apps, and websites

oai_token-act-pair · gpt-3.5-turbo

lists of items or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

imester

-0.87

izabeth

-0.83

cffffcc

-0.81

alysed

-0.76

icultural

-0.75

iring

-0.73

istical

-0.72

iscal

-0.71

duty

-0.70

emonic

-0.67

POSITIVE LOGITS

etc

1.04

 Chill

0.98

 Fract

0.98

Kik

0.98

 Naked

0.97

 Twisted

0.94

 Genie

0.94

 Sleeping

0.92

 Brav

0.92

 Scrib

0.92

Activations Density 0.237%

proper nouns related to entertainment and technology, such as movies, bands, apps, and websites

lists of items or entities

No Comments

No Known Activations