INDEX

Explanations

proper nouns, likely related to locations or names of people or organizations

oai_token-act-pair · gpt-3.5-turbo

coordinating conjunctions that connect ideas or phrases

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

>[

-0.79

Ĥİ

-0.76

ploy

-0.71

poke

-0.69

meet

-0.68

POST

-0.64

imil

-0.63

ibo

-0.63

meric

-0.63

physical

-0.62

POSITIVE LOGITS

 consequently

0.98

 thereby

0.98

 thus

0.95

 vice

0.93

 thence

0.93

 hence

0.91

 therefore

0.89

 assorted

0.88

 then

0.85

 subsequently

0.84

Activations Density 0.660%

proper nouns, likely related to locations or names of people or organizations

coordinating conjunctions that connect ideas or phrases

No Comments

No Known Activations