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

oran

-0.69

appropriately

-0.69

fen

-0.68

 kindred

-0.67

nikov

-0.66

atures

-0.64

ever

-0.63

 comr

-0.62

together

-0.62

gran

-0.62

POSITIVE LOGITS

AAAA

0.72

Score

0.69

itage

0.69

 Acceler

0.63

iott

0.62

uto

0.61

EEEE

0.60

=-=-

0.58

 Emanuel

0.57

 Wonders

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.