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

ench

-0.77

ischer

-0.69

','

-0.68

 Logged

-0.68

olas

-0.67

ells

-0.67

rehend

-0.67

icy

-0.66

rome

-0.66

eker

-0.65

POSITIVE LOGITS

 Impossible

0.74

ESCO

0.69

 Skype

0.68

ð

0.67

OTE

0.64

 Electoral

0.63

 concession

0.63

CGI

0.61

 Expend

0.60

 Soros

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.