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

yet

-0.76

 Hunger

-0.71

reetings

-0.67

 Magicka

-0.65

 forthcoming

-0.65

warts

-0.65

 Strikes

-0.60

 Winn

-0.59

Merit

-0.59

pport

-0.58

POSITIVE LOGITS

ufact

0.81

 arrang

0.76

 confir

0.75

assis

0.72

bilt

0.72

 occas

0.70

oln

0.70

undai

0.69

ership

0.69

 mosqu

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.