INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ateur

-0.91

atur

-0.72

aiden

-0.71

aurus

-0.68

ateurs

-0.68

warts

-0.68

urally

-0.67

rosso

-0.67

irit

-0.67

ummer

-0.66

POSITIVE LOGITS

leaf

0.69

 Agenda

0.69

 behalf

0.67

together

0.66

 leash

0.64

nep

0.61

 choke

0.59

 algorith

0.59

 bounty

0.59

 rigging

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.