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

jad

-0.66

 android

-0.65

 Cornel

-0.63

 Trout

-0.62

 Slate

-0.61

 Submit

-0.60

Pew

-0.60

 Lopez

-0.59

 Ventura

-0.58

 Suppose

-0.57

POSITIVE LOGITS

romeda

0.85

 horizont

0.82

ãĤ±

0.80

iasm

0.78

akespe

0.74

verty

0.73

 streng

0.71

ŃĶ

0.70

yss

0.69

ĸļ

0.68

Activations Density 0.001%

No Known Activations

This feature has no known activations.