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

service

-0.77

itu

-0.74

Lord

-0.73

Stock

-0.72

Rest

-0.69

Jim

-0.68

Operation

-0.67

Correction

-0.66

Trust

-0.65

----

-0.64

POSITIVE LOGITS

 horizont

0.76

aiman

0.68

pter

0.68

 opacity

0.68

hem

0.68

romy

0.67

avan

0.67

zag

0.66

cube

0.66

izont

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.