INDEX

Explanations

references to the term "busters."

oai_token-act-pair · gpt-3.5-turbo

references to the term "busters" 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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 temperament

-0.67

 horn

-0.64

ead

-0.64

 attribute

-0.63

 faint

-0.62

 convenience

-0.61

 trait

-0.60

ced

-0.60

 restriction

-0.60

End

-0.59

POSITIVE LOGITS

busters

5.01

buster

3.16

 Ghostbusters

1.18

 Buster

1.16

riots

1.02

 blockbuster

0.95

breakers

0.93

bots

0.89

roo

0.89

movie

0.88

Activations Density 0.025%

references to the term "busters."

references to the term "busters" in various contexts

No Comments

No Known Activations