INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inally

-0.82

mint

-0.81

thing

-0.74

ecd

-0.71

 pist

-0.70

straight

-0.70

atra

-0.70

cr

-0.67

reci

-0.67

 femin

-0.66

POSITIVE LOGITS

 Wolves

0.79

Ago

0.71

 Beasts

0.67

 Barrett

0.62

 Absent

0.62

 Towns

0.61

esville

0.61

Wil

0.60

 sheds

0.60

 deployments

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.