INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

reviewed

-0.72

 roundup

-0.69

"},

-0.69

Reason

-0.62

Buzz

-0.62

 Morton

-0.61

channelAvailability

-0.61

pak

-0.60

sta

-0.60

Pool

-0.59

POSITIVE LOGITS

ames

0.75

olas

0.75

 Aires

0.73

irements

0.73

awaru

0.72

raints

0.68

cedes

0.68

ensing

0.66

umo

0.66

imilar

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.