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

wild

-0.75

orio

-0.64

ME

-0.60

osi

-0.60

 overrun

-0.60

stellar

-0.60

 coordinating

-0.58

EEE

-0.58

theless

-0.58

iovascular

-0.57

POSITIVE LOGITS

 SetFontSize

0.79

iasis

0.71

igh

0.69

ograp

0.65

isations

0.64

 Tribune

0.64

 congratulations

0.64

isation

0.63

Topics

0.63

 squats

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.