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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ements

-0.73

beard

-0.73

tein

-0.70

onym

-0.70

 coli

-0.69

elist

-0.67

ethe

-0.66

Bey

-0.66

pedia

-0.66

assic

-0.64

POSITIVE LOGITS

 bonded

0.66

pun

0.65

 annex

0.63

 contracting

0.61

 plotted

0.59

 Roads

0.59

 compuls

0.58

vable

0.58

 destro

0.58

 Panel

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.