INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ochond

-0.78

them

-0.66

ffield

-0.63

 Longh

-0.62

spir

-0.61

abama

-0.61

 fielder

-0.60

 shit

-0.59

LSU

-0.58

mire

-0.57

POSITIVE LOGITS

rored

0.82

lasses

0.72

AGES

0.71

Else

0.68

 srfAttach

0.67

raft

0.66

reens

0.66

 Citation

0.65

ifts

0.64

terday

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.