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

ï¸

-0.79

Plot

-0.74

orph

-0.71

Spoiler

-0.70

KNOWN

-0.69

tesy

-0.69

mund

-0.68

nell

-0.68

PORT

-0.68

Flo

-0.67

POSITIVE LOGITS

enance

0.75

 farmers

0.72

 synd

0.67

 shar

0.67

 restore

0.66

 bolster

0.66

 restoring

0.64

 restoration

0.64

rers

0.63

 annihil

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.