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

 tack

-0.70

 repe

-0.68

 rese

-0.68

rupal

-0.64

 modelling

-0.63

 tame

-0.63

 route

-0.62

 bidder

-0.62

Mau

-0.61

 naming

-0.61

POSITIVE LOGITS

ombat

0.79

âĨ

0.76

 Remastered

0.74

anon

0.71

Mut

0.69

iewicz

0.68

Comment

0.67

omsky

0.67

Â¨

0.66

Footnote

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.