INDEX

Explanations

companies or organizations referenced as 'Co'

oai_token-act-pair · gpt-3.5-turbo

mentions of cooperation in various contexts

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

 IMAGES

-0.78

*/(

-0.69

Gor

-0.67

 RELEASE

-0.65

self

-0.63

 PEOPLE

-0.63

 Sabha

-0.62

nikov

-0.62

selves

-0.62

BEL

-0.61

POSITIVE LOGITS

Co

1.02

verage

0.98

agher

0.96

olly

0.93

aching

0.90

venant

0.90

operative

0.89

herent

0.89

lette

0.88

asters

0.85

Activations Density 0.010%

companies or organizations referenced as 'Co'

mentions of cooperation in various contexts

No Comments

No Known Activations