INDEX

Explanations

phrases related to comparison or contrasting different things

oai_token-act-pair · gpt-3.5-turbo

terms associated with health and medical risks

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

Hes

-0.56

Kah

-0.50

 Fields

-0.48

 Originally

-0.48

 Mesa

-0.47

 wisely

-0.47

 Reply

-0.46

 Posted

-0.46

↵Âł

-0.45

 Pelicans

-0.45

POSITIVE LOGITS

 imaginable

0.59

 unden

0.55

 undermin

0.55

jri

0.54

 charact

0.53

 behavi

0.53

ãĤ¢ãĥ«

0.52

ovan

0.50

 Flavoring

0.49

 characterization

0.49

Activations Density 2.044%

phrases related to comparison or contrasting different things

terms associated with health and medical risks

No Comments

No Known Activations