INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ogether

-0.83

untled

-0.76

soever

-0.75

iever

-0.75

uously

-0.73

mble

-0.72

icip

-0.71

essor

-0.71

paio

-0.71

inators

-0.71

POSITIVE LOGITS

burgh

0.69

0.66

Gw

0.64

lace

0.61

 Glen

0.59

 Memphis

0.58

ocr

0.57

gold

0.57

mem

0.56

 dens

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.