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

Rim

-0.74

gettable

-0.72

ushed

-0.69

orld

-0.64

Slot

-0.62

Elf

-0.62

aga

-0.60

ooked

-0.60

given

-0.59

zing

-0.59

POSITIVE LOGITS

ITS

0.77

 Mayhem

0.68

ļéĨĴ

0.67

 syll

0.65

anguage

0.65

 Retrieved

0.63

 aggregate

0.62

 GOODMAN

0.62

 youtube

0.62

KY

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.