INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

arij

-0.82

 Ajax

-0.72

xit

-0.72

 Carbuncle

-0.70

itiveness

-0.68

Coin

-0.65

 Petersen

-0.62

!/

-0.61

 Venezuel

-0.61

 Afric

-0.61

POSITIVE LOGITS

ailing

0.73

rollers

0.71

rogen

0.70

Gy

0.69

ails

0.69

 beam

0.65

 surg

0.64

sie

0.63

Girls

0.61

 minim

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.