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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enes

-0.86

intendent

-0.84

etsy

-0.80

nants

-0.79

grounds

-0.77

mens

-0.75

 McAuliffe

-0.74

itia

-0.73

uries

-0.71

ysc

-0.70

POSITIVE LOGITS

xual

0.71

ISH

0.69

LESS

0.65

_-

0.62

sax

0.61

 brill

0.61

 deaf

0.58

 saliva

0.57

 supplementary

0.57

AIR

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.