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

YC

-0.71

pring

-0.66

Xie

-0.66

bred

-0.66

rentices

-0.66

acial

-0.65

otle

-0.65

Yen

-0.65

elong

-0.64

zynski

-0.64

POSITIVE LOGITS

 prominently

0.82

Rated

0.70

 Splash

0.61

 Meter

0.60

 Notting

0.59

 parachute

0.59

ãĥĦ

0.58

 partitions

0.58

 lineback

0.56

Reviewer

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.