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

anyahu

-0.65

iger

-0.64

 Dolphin

-0.61

reement

-0.61

venth

-0.61

EStream

-0.60

alling

-0.59

 entitle

-0.59

 remedy

-0.58

¿½

-0.58

POSITIVE LOGITS

 Robbins

0.78

amac

0.73

ileaks

0.72

 Briggs

0.70

 Stephens

0.69

Brand

0.63

veland

0.63

 McKenzie

0.62

gas

0.62

 Brittany

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.