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

maximum

-0.73

 projection

-0.63

THR

-0.62

 Tort

-0.61

staff

-0.61

 alive

-0.60

IG

-0.60

 Excellence

-0.59

 retard

-0.58

YC

-0.58

POSITIVE LOGITS

iris

0.80

arine

0.77

orians

0.77

arios

0.77

itans

0.76

terday

0.76

oris

0.74

asley

0.69

ymes

0.69

 forgiven

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.