INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 blot

-0.90

emetery

-0.78

omy

-0.76

lip

-0.67

omic

-0.67

GBT

-0.66

omp

-0.66

é¾įå

-0.65

OME

-0.64

SEA

-0.62

POSITIVE LOGITS

 Mehran

0.73

inals

0.71

selves

0.69

 FranÃ§

0.68

assian

0.64

chev

0.62

 Stef

0.62

anguage

0.60

 Kazakh

0.60

#$#$

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.