INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

enic

-0.77

oha

-0.73

obe

-0.70

aming

-0.70

uggle

-0.64

amina

-0.63

born

-0.63

ague

-0.63

.ãĢį

-0.61

atri

-0.61

POSITIVE LOGITS

è¦ļéĨĴ

0.75

details

0.66

cfg

0.65

settings

0.64

 Instruments

0.64

models

0.63

iscons

0.62

sers

0.61

 Markets

0.61

Tes

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.