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

NetMessage

-0.77

@@

-0.74

cill

-0.72

...]

-0.70

aeda

-0.70

lator

-0.69

 Fixes

-0.67

Engineers

-0.65

bies

-0.64

âĢ¦]

-0.64

POSITIVE LOGITS

 shortest

0.67

Ath

0.62

 Huntington

0.62

 caption

0.58

 Slide

0.58

Pit

0.58

ngth

0.58

aceae

0.57

itas

0.57

 Beir

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.