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

 proport

-0.82

 obser

-0.80

 disse

-0.69

ciating

-0.69

 tentative

-0.66

ape

-0.65

 tremend

-0.65

 apes

-0.63

 miscon

-0.62

 toppled

-0.61

POSITIVE LOGITS

ews

0.89

 Profile

0.82

âĢİ

0.80

quickShipAvailable

0.76

atri

0.74

 âĢİ

0.72

rawdownloadcloneembedreportprint

0.70

roth

0.69

agher

0.67

antry

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.