INDEX

Explanations

information or content related to legal documents and texts

oai_token-act-pair · gpt-3.5-turbo

referential phrases indicating the presence of a subject in a statement or report

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

 behav

-0.67

 Ging

-0.63

 Antar

-0.59

 reins

-0.59

 Meteor

-0.57

 latter

-0.56

 thereto

-0.55

 Tacoma

-0.55

 later

-0.55

 Gund

-0.54

POSITIVE LOGITS

[/

0.91

resa

0.78

[+

0.72

 guiName

0.72

Donald

0.71

[/

0.71

orem

0.71

chieve

0.70

auri

0.69

ablishment

0.69

Activations Density 0.353%

information or content related to legal documents and texts

referential phrases indicating the presence of a subject in a statement or report

No Comments

No Known Activations