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

 indu

-0.67

Ky

-0.65

oes

-0.64

 resc

-0.62

 supplies

-0.62

ieth

-0.61

Sik

-0.61

 Zoro

-0.60

 appra

-0.60

 prophes

-0.59

POSITIVE LOGITS

iral

0.84

 largeDownload

0.83

tumblr

0.79

cean

0.75

adden

0.71

SQL

0.71

brow

0.70

zyme

0.70

ahoo

0.69

ety

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.