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

eous

-0.67

 grain

-0.62

 Useful

-0.60

 Pros

-0.59

*)

-0.59

ares

-0.59

*.

-0.58

?]

-0.58

 Singh

-0.58

reau

-0.58

POSITIVE LOGITS

naires

0.77

ilater

0.74

Anim

0.68

CLOSE

0.68

lain

0.67

ansas

0.67

ria

0.65

uador

0.64

aurus

0.64

occupied

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.