INDEX

Explanations

combinations or incorporations of different elements into a unified whole

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the integration or combination of multiple elements into a cohesive whole

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ãĥ©ãĥ³

-0.69

awa

-0.62

ander

-0.62

esan

-0.61

ffect

-0.60

lean

-0.58

ãĥĥãĤ¯

-0.58

lash

-0.58

rued

-0.58

mark

-0.57

POSITIVE LOGITS

one

1.02

 cohesive

0.99

ONE

0.90

 single

0.90

 coherent

0.83

 manageable

0.83

 harmony

0.82

 common

0.82

 unison

0.81

 singular

0.81

Activations Density 0.241%

combinations or incorporations of different elements into a unified whole

phrases indicating the integration or combination of multiple elements into a cohesive whole

No Comments

No Known Activations