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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ËĪ

-0.90

ãĥ¼ãĥĨãĤ£

-0.75

 actionGroup

-0.74

ãĥ£

-0.73

urion

-0.69

ãĥĥãĥī

-0.68

Ùİ

-0.67

oun

-0.65

sburg

-0.65

orer

-0.64

POSITIVE LOGITS

METHOD

0.72

Closure

0.69

 Anthrop

0.68

Course

0.67

Ros

0.66

Ingredients

0.65

Methods

0.65

rition

0.63

Continue

0.62

Results

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.