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

angelo

-0.68

akeru

-0.66

ctory

-0.64

opol

-0.63

 guiActiveUn

-0.62

 appeal

-0.61

urations

-0.60

 antid

-0.60

isure

-0.60

res

-0.60

POSITIVE LOGITS

ð

0.73

PK

0.72

peg

0.70

laws

0.69

ccording

0.68

roller

0.66

iaries

0.66

headed

0.65

($)

0.65

Õ

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.