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

eeper

-0.74

affe

-0.74

aii

-0.72

enegger

-0.71

rament

-0.70

epad

-0.68

urally

-0.68

ebus

-0.67

natureconservancy

-0.64

odox

-0.63

POSITIVE LOGITS

 Bulg

0.75

 Warsaw

0.72

 pods

0.63

 Notting

0.63

 Cecil

0.62

 Scully

0.61

flix

0.61

 Baltic

0.60

Wee

0.60

 Kafka

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.