INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 face

-0.65

é¾

-0.64

 beats

-0.61

 inputs

-0.60

 square

-0.60

IGHT

-0.60

 drills

-0.60

 checkpoints

-0.59

 overlap

-0.57

 Euros

-0.56

POSITIVE LOGITS

ournal

0.91

pherd

0.81

terness

0.81

gerald

0.78

hement

0.77

sembly

0.75

zie

0.74

lahoma

0.72

ikini

0.72

ntil

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.