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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 resil

-0.88

 conclud

-0.76

sth

-0.74

ttes

-0.74

 experien

-0.74

iotics

-0.69

 Bucc

-0.68

ullivan

-0.67

 concess

-0.67

 Survivors

-0.67

POSITIVE LOGITS

Org

0.68

aire

0.67

hung

0.62

 Eisen

0.62

fed

0.61

oa

0.60

ayer

0.60

SPONSORED

0.58

Pin

0.58

Reading

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.