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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nen

-0.81

ãĤ¦

-0.70

ramid

-0.69

oman

-0.66

 Weasley

-0.65

bourne

-0.63

lete

-0.62

grave

-0.61

 Pants

-0.61

 Patron

-0.61

POSITIVE LOGITS

up

1.17

UP

0.90

Up

0.85

 guiActiveUn

0.81

Up

0.77

up

0.75

usc

0.71

UP

0.69

arser

0.68

 manag

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.