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

idon

-0.82

ilogy

-0.79

halla

-0.79

arella

-0.78

acea

-0.77

PsyNetMessage

-0.76

ÃŃa

-0.76

renheit

-0.74

 resid

-0.74

ertodd

-0.73

POSITIVE LOGITS

 upright

0.71

cms

0.66

 pressing

0.61

 harmless

0.61

 juvenile

0.61

0.60

 Devon

0.59

 shoulder

0.58

 improper

0.56

 intrusive

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.