INDEX

Explanations

statements or reports from official sources

oai_token-act-pair · gpt-3.5-turbo

references to statements and announcements attributed to various entities or sources

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

wick

-0.74

 carbohyd

-0.73

chwitz

-0.72

endish

-0.71

gradation

-0.69

lied

-0.68

merce

-0.68

ighth

-0.67

faced

-0.67

ritional

-0.66

POSITIVE LOGITS

 afar

1.10

 inside

0.86

 whence

0.86

 abroad

0.81

 within

0.74

 thence

0.74

 Inside

0.71

 Rabbi

0.70

 atop

0.69

 Washington

0.68

Activations Density 0.099%

statements or reports from official sources

references to statements and announcements attributed to various entities or sources

No Comments

No Known Activations