INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Offline

-0.86

atz

-0.80

orem

-0.73

AAAA

-0.73

IED

-0.72

ioxide

-0.71

Log

-0.68

erent

-0.67

emo

-0.66

obal

-0.65

POSITIVE LOGITS

child

0.68

Kes

0.66

 barriers

0.63

inement

0.61

hur

0.60

 Finn

0.60

 barrier

0.59

aviour

0.59

lins

0.58

 stretches

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.