INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

issues

-0.78

çĶ°

-0.77

itus

-0.76

xual

-0.72

ota

-0.71

poral

-0.68

dan

-0.67

ega

-0.67

ochond

-0.67

omo

-0.66

POSITIVE LOGITS

 tipped

0.69

 guaranteed

0.64

 broccoli

0.63

 glyphosate

0.62

 agric

0.62

 protected

0.62

osate

0.60

 Roundup

0.58

 Monsanto

0.57

 alarmed

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.