INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

arted

-0.71

 hawk

-0.67

 infect

-0.60

alsh

-0.60

LEASE

-0.59

ħĭ

-0.59

 devils

-0.59

 radiation

-0.59

ali

-0.58

 dwindling

-0.58

POSITIVE LOGITS

aic

0.84

 Rodney

0.69

Balance

0.67

olean

0.67

issan

0.66

oby

0.64

ophysical

0.64

orius

0.63

ocom

0.62

 Credit

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.