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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ships

-0.78

ship

-0.76

helm

-0.68

jen

-0.66

sea

-0.65

edit

-0.65

aird

-0.62

lord

-0.62

 Cthulhu

-0.62

kas

-0.61

POSITIVE LOGITS

ĸļ

0.92

 proport

0.81

MSG

0.74

ģ«

0.67

¥ŀ

0.63

ument

0.60

 syll

0.60

oppable

0.60

erved

0.59

awaru

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.