INDEX

Explanations

phrases indicating belief or perception

oai_token-act-pair · gpt-3.5-turbo

the verb "to be" in various forms and contexts in the text

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

Scroll

-0.80

pedia

-0.74

 Salon

-0.73

 Verge

-0.69

ilst

-0.66

 Gaul

-0.66

 Britons

-0.66

 HuffPost

-0.63

 resorts

-0.63

 Investigative

-0.60

POSITIVE LOGITS

 gonna

1.29

ELF

0.94

 funny

0.94

 okay

0.89

 going

0.85

ok

0.85

 happening

0.84

leeve

0.81

 wrong

0.80

 gotta

0.79

Activations Density 0.216%

phrases indicating belief or perception

the verb "to be" in various forms and contexts in the text

No Comments

No Known Activations