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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.70

pill

-0.69

rawdownloadcloneembedreportprint

-0.69

poons

-0.68

olson

-0.67

 Chow

-0.67

wald

-0.66

poon

-0.65

giene

-0.65

 averages

-0.64

POSITIVE LOGITS

SCP

0.76

ERN

0.71

 activated

0.68

)]

0.67

 affirmative

0.66

]);

0.65

)].

0.64

DragonMagazine

0.64

 certs

0.63

tarian

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.