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

lled

-0.73

ng

-0.70

tm

-0.63

lly

-0.62

mb

-0.61

windows

-0.61

----------

-0.60

bed

-0.60

plex

-0.59

âĶ

-0.59

POSITIVE LOGITS

uyomi

0.84

 veter

0.78

ebus

0.74

0.73

livion

0.71

 equival

0.69

 mint

0.68

dinand

0.66

chwitz

0.66

 Baal

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.