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

efe

-0.70

 shake

-0.68

 negotiate

-0.68

usterity

-0.63

 Constantin

-0.62

hr

-0.61

elight

-0.61

rake

-0.60

union

-0.60

raid

-0.59

POSITIVE LOGITS

SPONSORED

0.83

mentioned

0.81

amples

0.75

ILA

0.72

alist

0.71

doi

0.70

lined

0.67

criptions

0.64

vantage

0.63

////////////////////////////////

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.