INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

hered

-0.71

isher

-0.67

opher

-0.65

jen

-0.64

ername

-0.63

ophers

-0.63

ayn

-0.62

mask

-0.61

omen

-0.61

arij

-0.61

POSITIVE LOGITS

 Eggs

0.74

venants

0.67

 DISTR

0.66

psey

0.65

 Minutes

0.64

reditary

0.63

Ĥª

0.62

ilibrium

0.62

å£«

0.61

cffffcc

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.