INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

udeb

-0.84

awaru

-0.68

assetsadobe

-0.65

 Expand

-0.62

 yarn

-0.60

"},"

-0.59

 productions

-0.58

[|

-0.58

 Concepts

-0.58

biz

-0.58

POSITIVE LOGITS

 Stead

0.79

 Siem

0.78

orr

0.71

icter

0.69

Â·Â·

0.69

 Titus

0.68

ificate

0.67

abouts

0.65

 Mississ

0.63

gins

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.