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

amd

-0.71

endor

-0.71

ition

-0.70

hiba

-0.70

ety

-0.70

atum

-0.69

atical

-0.69

elin

-0.66

ilus

-0.66

estone

-0.66

POSITIVE LOGITS

 space

0.78

 Berks

0.75

GHC

0.71

 overcrowd

0.67

orship

0.66

 nodd

0.65

 popping

0.64

 spaces

0.64

 cricket

0.64

ellar

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.