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

 Sending

-0.66

 retweet

-0.65

_>

-0.62

Rey

-0.60

 Grab

-0.59

 Dancing

-0.58

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.56

ª

-0.56

Ney

-0.56

 Sterling

-0.55

POSITIVE LOGITS

outed

0.79

lon

0.77

amn

0.74

ickets

0.73

osponsors

0.73

heit

0.71

authorized

0.71

lich

0.71

icket

0.71

adel

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.