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

won

-0.71

 encountering

-0.71

 suspending

-0.70

 celebrated

-0.68

 recogn

-0.67

 appreciated

-0.67

 cancell

-0.66

 mastered

-0.64

ilty

-0.64

 anchored

-0.64

POSITIVE LOGITS

 ILCS

0.81

 Homeless

0.78

Harris

0.70

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

0.69

,,,,,,,,

0.68

Ø¹

0.68

ODUCT

0.67

ridor

0.67

nex

0.67

Ø³

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.