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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agher

-0.76

uits

-0.72

wcsstore

-0.69

atta

-0.65

>(

-0.64

ureen

-0.63

ubes

-0.62

ello

-0.61

ulla

-0.61

 amen

-0.61

POSITIVE LOGITS

ç·

0.64

ãĥī

0.63

FK

0.60

 Dunk

0.60

ï¸

0.59

Amazon

0.59

Buzz

0.58

advertisement

0.58

dra

0.57

 Sutherland

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.