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

FIA

-0.76

theless

-0.74

SDL

-0.71

maid

-0.63

 batter

-0.63

 Trafford

-0.61

 Atari

-0.61

bah

-0.60

ts

-0.59

HT

-0.58

POSITIVE LOGITS

spr

0.73

arthed

0.73

ONSORED

0.70

orter

0.69

Reward

0.68

rant

0.67

"]=>

0.66

owsky

0.65

endas

0.64

orters

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.