INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ikuman

-1.05

ãĥ¼ãĥĨãĤ£

-0.84

 livest

-0.84

 encount

-0.71

psey

-0.68

Thumbnail

-0.67

 recognise

-0.65

abetic

-0.65

Downloadha

-0.61

PDATE

-0.61

POSITIVE LOGITS

0.96

,...

0.76

loo

0.76

udi

0.74

for

0.71

 ASAP

0.67

 Frank

0.67

 Gupta

0.66

,,

0.65

 Indies

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.