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

azeera

-0.80

arma

-0.80

gee

-0.79

ledged

-0.77

 flank

-0.71

inav

-0.70

ritz

-0.69

stadt

-0.69

ammy

-0.63

zeb

-0.63

POSITIVE LOGITS

 Vinyl

0.81

ilts

0.76

 Plaint

0.73

 FANTASY

0.72

 Clause

0.70

 Learns

0.66

MAS

0.64

frames

0.64

ypes

0.64

âĸº

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.