INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ignty

-0.77

araoh

-0.68

etheless

-0.63

Ob

-0.63

verning

-0.61

asty

-0.61

enger

-0.60

angled

-0.60

fal

-0.59

inating

-0.58

POSITIVE LOGITS

hop

0.66

 Rollins

0.66

ORN

0.65

went

0.65

razil

0.64

pmwiki

0.64

HOU

0.63

iatus

0.63

uminati

0.62

haus

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.