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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Helpful

-0.72

ascript

-0.66

ugu

-0.65

 Thoughts

-0.61

 Dear

-0.60

________________________________________________________________

-0.57

 Plain

-0.56

agascar

-0.56

 mathemat

-0.56

Asc

-0.56

POSITIVE LOGITS

LAN

0.79

LOD

0.72

requires

0.70

spir

0.70

 Lerner

0.67

urion

0.67

wake

0.67

lore

0.67

abwe

0.66

?),

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.