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

erers

-0.79

 Signs

-0.73

 Aware

-0.70

æ©

-0.69

 Volunte

-0.68

GROUP

-0.67

Tea

-0.67

anism

-0.67

encer

-0.65

 Byron

-0.65

POSITIVE LOGITS

 expiration

0.79

riage

0.71

 zipper

0.70

 trademark

0.67

 countdown

0.67

 violation

0.66

 retrieval

0.65

ado

0.65

 date

0.64

 search

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.