INDEX

Explanations

specific institutions or locations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "the."

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

 besides

-0.82

Enjoy

-0.75

 asleep

-0.67

 innovate

-0.66

 peanuts

-0.65

 anecd

-0.64

aji

-0.63

gpu

-0.62

omever

-0.62

 furthermore

-0.62

POSITIVE LOGITS

 Philippines

0.99

 Netherlands

0.95

 latter

0.95

 same

0.93

 United

0.91

 aftermath

0.89

 Gulf

0.87

 infamous

0.84

 aforementioned

0.83

 National

0.82

Activations Density 0.702%

specific institutions or locations

instances of the word "the."

No Comments

No Known Activations