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

 Piercing

-0.72

dx

-0.71

ences

-0.70

ioxide

-0.70

urities

-0.67

 sqor

-0.67

 Races

-0.67

erning

-0.66

INESS

-0.66

ãĥ¯

-0.66

POSITIVE LOGITS

 managers

0.72

 management

0.72

 rent

0.69

 babys

0.68

 rede

0.67

 park

0.67

 manage

0.65

 decor

0.64

 pageant

0.64

 rake

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.