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

çīĪ

-0.80

uckles

-0.77

rha

-0.69

 impacted

-0.68

irit

-0.66

ickr

-0.65

dden

-0.65

 affected

-0.65

ħĭ

-0.63

itures

-0.61

POSITIVE LOGITS

votes

0.77

atur

0.67

orian

0.65

ween

0.64

stable

0.64

mat

0.63

acci

0.63

athe

0.63

til

0.62

pite

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.