INDEX

Explanations

new elements or concepts within a context

oai_token-act-pair · gpt-3.5-turbo

references to dualities or comparisons involving the concept of "both."

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

 Parables

-0.62

Ļ½

-0.57

 Ratio

-0.57

 extrad

-0.57

 Amen

-0.56

 Beir

-0.56

 Genius

-0.55

 Trace

-0.55

 pseudonym

-0.54

ossier

-0.54

POSITIVE LOGITS

nor

0.84

 sexes

0.81

"]=>

0.80

fml

0.69

 statically

0.66

selves

0.65

ELY

0.64

_>

0.63

idge

0.63

eers

0.63

Activations Density 0.208%

new elements or concepts within a context

references to dualities or comparisons involving the concept of "both."

No Comments

No Known Activations