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

lease

-0.74

Ohio

-0.68

urd

-0.68

orsche

-0.66

bp

-0.66

Asset

-0.65

Commercial

-0.65

arij

-0.64

:(

-0.64

imp

-0.64

POSITIVE LOGITS

 thus

0.84

 therefore

0.79

 consequently

0.75

 thereby

0.74

etc

0.71

 then

0.71

 hence

0.69

 obey

0.67

 thence

0.66

 vice

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.