INDEX

Explanations

words related to publication and publishers

oai_token-act-pair · gpt-3.5-turbo

phrases attributing authorship or publication to works

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

xual

-0.83

hene

-0.82

uality

-0.81

olit

-0.78

itialized

-0.78

itiveness

-0.76

enery

-0.74

iated

-0.74

 manif

-0.74

earchers

-0.73

POSITIVE LOGITS

Viz

0.82

 Wizards

0.80

 Nationwide

0.79

 Warner

0.79

 Lions

0.77

Nex

0.75

 Sony

0.75

Wad

0.75

 Saban

0.75

 Univ

0.75

Activations Density 0.118%

words related to publication and publishers

phrases attributing authorship or publication to works

No Comments

No Known Activations