INDEX

Explanations

words or phrases that are being described or defined

oai_token-act-pair · gpt-3.5-turbo

instances of the word "characterized" and its variations in different 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

 Â£

-0.76

 maths

-0.76

 Nasa

-0.72

Tes

-0.70

Kiw

-0.67

 extra

-0.66

 keeper

-0.65

 practise

-0.64

 Gods

-0.63

RAF

-0.63

POSITIVE LOGITS

 characterized

3.62

 characterize

2.33

 termed

1.54

 categorized

1.43

 summarized

1.38

 defined

1.36

 characterization

1.35

 portrayed

1.31

 fueled

1.26

 described

1.25

Activations Density 0.013%

words or phrases that are being described or defined

instances of the word "characterized" and its variations in different contexts

No Comments

No Known Activations