INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ariat

-0.73

puters

-0.67

 Robotics

-0.67

ysis

-0.64

 stationary

-0.63

eks

-0.62

 stagn

-0.60

icit

-0.60

arial

-0.59

oke

-0.59

POSITIVE LOGITS

amaz

0.73

detail

0.70

Lind

0.66

aston

0.64

âĨĳ

0.64

)=(

0.64

":"","

0.63

oshenko

0.63

intent

0.62

 Albion

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.