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

 Reincarn

-0.74

 Johns

-0.71

 Quest

-0.67

 Bain

-0.66

 Rapt

-0.66

Rox

-0.64

 Medal

-0.64

 Catalyst

-0.64

 Correct

-0.64

 Secret

-0.63

POSITIVE LOGITS

uana

0.84

tten

0.81

senal

0.78

EStream

0.76

rafted

0.74

icts

0.74

nants

0.74

=]

0.72

escription

0.72

ondon

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.