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

onica

-0.73

Sax

-0.61

 actionGroup

-0.61

 affirmative

-0.59

ancial

-0.58

 folding

-0.57

 Raleigh

-0.55

stead

-0.55

 unison

-0.55

rye

-0.54

POSITIVE LOGITS

ESE

0.74

pec

0.72

éģ

0.69

Oracle

0.68

isman

0.67

DATA

0.67

Pand

0.67

steen

0.66

AP

0.66

pherd

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.