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

mination

-0.69

ificent

-0.64

 noise

-0.62

umph

-0.62

 negatives

-0.61

ãĤĭ

-0.61

ainment

-0.61

uitous

-0.59

akespe

-0.59

Neg

-0.58

POSITIVE LOGITS

 laun

0.76

Charg

0.75

 psychiat

0.75

cdn

0.71

0.67

 Neuroscience

0.65

ĸļ

0.64

yss

0.62

astern

0.62

 Nanto

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.