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

 mischief

-0.64

Hod

-0.63

PD

-0.62

 filter

-0.61

TAG

-0.61

VPN

-0.60

use

-0.59

 lurking

-0.58

 store

-0.58

ãĥĥãĥĪ

-0.57

POSITIVE LOGITS

 sidx

0.77

reprene

0.75

ocrats

0.71

 guiIcon

0.70

icion

0.70

rists

0.70

oshenko

0.69

aturdays

0.68

 Prosecut

0.67

ctory

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.