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

vernment

-0.75

inals

-0.70

entin

-0.66

wr

-0.66

roman

-0.65

population

-0.64

minster

-0.64

inates

-0.64

 plotting

-0.63

 neighb

-0.63

POSITIVE LOGITS

 Tasman

0.73

 Advocate

0.70

ARGET

0.69

DeL

0.69

nel

0.65

KK

0.65

enge

0.65

acht

0.64

 Wikileaks

0.63

eto

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.