INDEX

Explanations

instances where events are described in relation to time

oai_token-act-pair · gpt-3.5-turbo

phrases indicating time and frequency

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

hor

-0.50

oya

-0.49

Oregon

-0.49

hop

-0.48

 Prosecut

-0.47

-)

-0.46

intention

-0.46

]'

-0.45

)'

-0.45

hal

-0.44

POSITIVE LOGITS

 fructose

0.53

*/(

0.51

ielding

0.51

iatrics

0.48

widget

0.47

 favour

0.46

 frightening

0.46

 undet

0.45

 disguise

0.45

 unsuspecting

0.44

Activations Density 0.833%

instances where events are described in relation to time

phrases indicating time and frequency

No Comments

No Known Activations