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

TPPStreamerBot

-0.74

 constitu

-0.72

DAC

-0.70

ORPG

-0.68

 contracted

-0.66

CIS

-0.63

acco

-0.61

 Neph

-0.61

 ancestor

-0.59

 slightest

-0.58

POSITIVE LOGITS

rying

0.70

 Lanka

0.68

 Folk

0.67

elling

0.65

illet

0.64

gew

0.64

rief

0.64

olo

0.63

ricks

0.62

otine

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.