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

abouts

-0.69

 Guilty

-0.67

Hos

-0.67

 hypers

-0.67

 boarding

-0.66

HR

-0.66

zig

-0.65

ploy

-0.63

ammy

-0.62

uct

-0.62

POSITIVE LOGITS

resents

0.81

Quantity

0.76

Methods

0.76

Future

0.75

values

0.74

Results

0.71

Series

0.69

rien

0.69

fm

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.