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

ãĥ¼ãĤ¯

-0.77

 weap

-0.68

 Prev

-0.67

MpServer

-0.64

 subdu

-0.61

 behavi

-0.59

 challeng

-0.59

 relief

-0.58

 enhancement

-0.58

 modulation

-0.57

POSITIVE LOGITS

berra

0.82

ype

0.82

ete

0.78

milo

0.77

ethical

0.76

agne

0.75

TED

0.73

tle

0.72

lass

0.71

 Virtue

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.