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

essee

-0.81

itution

-0.69

Cart

-0.68

 Cart

-0.66

ecd

-0.65

ittee

-0.64

ertodd

-0.63

abor

-0.62

stown

-0.62

otyp

-0.61

POSITIVE LOGITS

vet

0.70

 facebook

0.66

zac

0.65

 wills

0.64

haz

0.63

untu

0.63

 Tsukuyomi

0.63

zan

0.63

 fortunes

0.62

 reckon

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.