INDEX

Explanations

words related to things becoming more intense or serious

oai_token-act-pair · gpt-3.5-turbo

words related to processes or states of becoming or increasing in intensity

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 independently

-0.66

é¾

-0.63

Ty

-0.61

DEV

-0.59

 legally

-0.57

color

-0.56

cffffcc

-0.56

 unsuccessfully

-0.56

 solo

-0.56

zip

-0.55

POSITIVE LOGITS

ening

4.42

ened

3.09

eners

2.39

ener

2.29

ens

1.98

enment

1.52

ENS

1.38

eness

1.30

thening

1.29

en

1.28

Activations Density 0.006%

words related to things becoming more intense or serious

words related to processes or states of becoming or increasing in intensity

No Comments

No Known Activations