INDEX

Explanations

references to reports or studies specifically mentioning different organizations or entities

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

itars

-0.79

 besides

-0.75

 rever

-0.73

plings

-0.73

uality

-0.71

 eleph

-0.69

assed

-0.68

coins

-0.68

pler

-0.67

obyl

-0.66

POSITIVE LOGITS

 Department

1.14

 United

1.14

 National

1.13

 Ministry

1.10

 Associated

1.06

 ACLU

1.00

 Association

1.00

 Federal

0.98

 latter

0.97

 International

0.97

Activations Density 0.385%

references to reports or studies specifically mentioning different organizations or entities

No Comments

No Known Activations