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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

WER

-0.68

achine

-0.66

agent

-0.66

SON

-0.63

States

-0.60

 nomine

-0.59

SG

-0.59

ribute

-0.58

 Leban

-0.57

hr

-0.57

POSITIVE LOGITS

utics

0.72

nces

0.72

ophob

0.70

uras

0.69

tics

0.69

upuncture

0.65

aign

0.65

Ĥª

0.64

 Catholics

0.62

Haz

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.