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

warning

-0.71

yz

-0.71

wal

-0.67

="#

-0.65

Companies

-0.65

stick

-0.64

dirty

-0.63

 divest

-0.63

_.

-0.63

footed

-0.61

POSITIVE LOGITS

 Aval

0.75

 Mercer

0.73

âĶĢâĶĢâĶĢâĶĢ

0.70

 Proud

0.70

 enthusi

0.70

Ã°

0.68

 breath

0.67

ience

0.67

otti

0.66

unction

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.