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

soever

-0.73

 confir

-0.68

 experien

-0.65

-0.62

osphere

-0.62

 collateral

-0.61

 elig

-0.61

haps

-0.61

ilk

-0.58

phen

-0.57

POSITIVE LOGITS

 Aires

0.73

rael

0.73

nea

0.69

hift

0.66

 Adin

0.66

tail

0.65

Sax

0.64

 Rhod

0.64

UTION

0.64

 Reef

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.