INDEX

Explanations

phrases related to product or technology launches

oai_token-act-pair · gpt-3.5-turbo

instances of the word "introducing" and its variations related to the introduction of new topics or products

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

 Immunity

-0.78

 Mara

-0.70

 Frankfurt

-0.68

 sleeper

-0.65

[+

-0.65

 prevailing

-0.64

 Belt

-0.63

 bearer

-0.62

 Fifty

-0.62

 amusement

-0.62

POSITIVE LOGITS

ctory

1.36

ctions

1.34

pta

1.08

cing

1.06

ction

0.97

ce

0.96

cÃ©

0.92

ctive

0.91

cin

0.90

acia

0.90

Activations Density 0.033%

phrases related to product or technology launches

instances of the word "introducing" and its variations related to the introduction of new topics or products

No Comments

No Known Activations