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

 Medals

-0.83

 Medal

-0.77

PLA

-0.76

ORED

-0.75

Narr

-0.73

 rall

-0.73

 Seym

-0.72

ãĤ´

-0.68

ieth

-0.67

 itiner

-0.67

POSITIVE LOGITS

ktop

0.82

utterstock

0.67

 Sigma

0.66

apt

0.65

 likely

0.63

 probably

0.62

 tether

0.62

 leverage

0.61

illance

0.60

 infect

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.