INDEX

Explanations

verbs related to the action of applying something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "applied" 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

buster

-0.66

score

-0.66

 Throne

-0.64

 Reneg

-0.64

ghazi

-0.63

dest

-0.63

izen

-0.63

Rum

-0.62

watch

-0.62

flame

-0.61

POSITIVE LOGITS

 applied

3.73

 applying

2.13

 apply

2.11

 Applied

1.87

 applies

1.86

 Apply

1.83

apply

1.74

Apply

1.73

 application

1.44

 applicable

1.36

Activations Density 0.009%

verbs related to the action of applying something

instances of the word "applied" in various contexts

No Comments

No Known Activations