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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vae

-0.66

---------

-0.66

-->

-0.65

$$

-0.64

plays

-0.64

ittle

-0.64

creen

-0.63

xs

-0.62

hens

-0.62

————

-0.61

POSITIVE LOGITS

 conflic

0.76

senal

0.73

EStream

0.68

 Ambro

0.66

artifacts

0.65

ufact

0.62

 enthusi

0.61

 Dise

0.61

 glim

0.61

 reluct

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.