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

 respectively

-0.71

Bak

-0.64

DEM

-0.61

 misled

-0.59

 derived

-0.59

 inspections

-0.58

AGES

-0.57

 arising

-0.56

 balls

-0.56

 apologise

-0.56

POSITIVE LOGITS

tumblr

0.83

Offline

0.82

uras

0.78

DonaldTrump

0.75

natureconservancy

0.74

metic

0.74

reon

0.73

atform

0.73

tsy

0.72

teness

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.