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

ital

-0.81

morrow

-0.80

Cod

-0.79

Reviewed

-0.79

speak

-0.78

cyclopedia

-0.77

acca

-0.76

lique

-0.76

itled

-0.75

Blog

-0.71

POSITIVE LOGITS

 Survivor

0.70

 elimination

0.68

CG

0.64

 loser

0.63

 Tribal

0.62

osate

0.61

IPM

0.61

Miz

0.60

 Removal

0.60

 moves

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.