INDEX

Explanations

words related to personal or informal opinions and expressions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating uncertainty or doubt

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

emale

-0.66

anwhile

-0.56

ä»£

-0.56

senal

-0.54

 eleph

-0.52

vertisement

-0.52

culosis

-0.50

artney

-0.50

milo

-0.49

cliffe

-0.49

POSITIVE LOGITS

 Naruto

0.46

 REALLY

0.45

 Spoiler

0.44

pmwiki

0.44

)--

0.44

 spoiler

0.44

 messed

0.44

 âĢº

0.42

 weird

0.42

 kinda

0.42

Activations Density 2.175%

words related to personal or informal opinions and expressions

phrases indicating uncertainty or doubt

No Comments

No Known Activations