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

ilater

-0.74

tails

-0.67

jaw

-0.64

ammy

-0.64

hon

-0.61

 Peng

-0.59

uador

-0.59

PM

-0.58

uber

-0.58

anders

-0.58

POSITIVE LOGITS

ablishment

0.73

ENA

0.71

ãĤ´ãĥ³

0.70

tsky

0.70

 CODE

0.69

iece

0.66

 suffice

0.65

 pleading

0.65

artment

0.64

 Despair

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.