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

acia

-0.81

igree

-0.77

ici

-0.74

inqu

-0.73

icity

-0.71

eday

-0.70

tera

-0.68

abase

-0.68

elaide

-0.66

 horizont

-0.65

POSITIVE LOGITS

OTHER

0.67

 guest

0.64

iasis

0.63

 john

0.63

keeper

0.63

 Instruction

0.59

 author

0.59

 inherit

0.59

 photoc

0.58

 latex

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.