INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

CLASSIFIED

-0.75

cens

-0.69

[+

-0.69

":"/

-0.67

 Folk

-0.66

 Again

-0.65

 Brush

-0.64

itiveness

-0.64

RG

-0.63

LCS

-0.63

POSITIVE LOGITS

omnia

0.73

ulty

0.72

 weary

0.67

hof

0.65

 ages

0.65

hung

0.64

odo

0.64

orians

0.63

ories

0.62

axies

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.