INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

SEC

-0.66

uld

-0.66

ysis

-0.65

erate

-0.63

station

-0.63

ilogy

-0.62

reb

-0.61

istar

-0.61

ECD

-0.61

otine

-0.61

POSITIVE LOGITS

ners

0.70

overs

0.68

atche

0.66

 ACTIONS

0.66

soType

0.66

aughs

0.63

offs

0.63

avored

0.62

phasis

0.62

ees

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.