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

 Rest

-0.67

 bind

-0.63

 Prev

-0.61

 Walt

-0.61

 Herz

-0.60

 most

-0.58

REM

-0.58

 Twisted

-0.58

 fore

-0.58

 Prom

-0.58

POSITIVE LOGITS

who

0.95

whose

0.81

sonian

0.75

who

0.74

nesota

0.71

CTV

0.70

WHO

0.69

esses

0.69

CVE

0.68

ingen

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.