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

 Seym

-0.77

vati

-0.75

afort

-0.71

 cassette

-0.66

raints

-0.65

 Archdemon

-0.65

 counselor

-0.63

 Plymouth

-0.63

eredith

-0.61

 bead

-0.61

POSITIVE LOGITS

votes

0.74

cause

0.68

UID

0.68

itiveness

0.68

intention

0.66

epad

0.65

omatic

0.65

Bird

0.64

odo

0.64

 guiActiveUnfocused

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.