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

 Glob

-0.75

 Theft

-0.73

loader

-0.69

 Statements

-0.68

pict

-0.68

lasses

-0.68

Component

-0.67

cycl

-0.67

 Airl

-0.65

odor

-0.65

POSITIVE LOGITS

ufact

0.75

 disse

0.70

 resid

0.66

 coerc

0.65

reditary

0.65

qqa

0.63

ascus

0.63

 trou

0.62

 metic

0.62

 sued

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.