INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

emark

-0.79

alian

-0.79

AAAAAAAA

-0.77

enburg

-0.73

ework

-0.73

agu

-0.73

pse

-0.72

 bankrupt

-0.69

agall

-0.69

mort

-0.69

POSITIVE LOGITS

 Ashton

0.79

 Pearce

0.73

Kit

0.72

 Catalyst

0.70

 Barker

0.66

 Scope

0.65

 Jensen

0.65

 Summit

0.64

 Emirates

0.64

 Extensions

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.