INDEX

Explanations

phrases related to comparison or contrast

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate assessments or evaluations of situations

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

eteenth

-0.83

DAQ

-0.82

aughtered

-0.81

oward

-0.77

cedented

-0.75

atform

-0.75

aic

-0.74

asca

-0.73

akh

-0.71

soever

-0.71

POSITIVE LOGITS

 wanting

1.53

 liking

1.45

 hating

1.40

 having

1.39

 needing

1.38

 letting

1.38

 forgetting

1.36

 knowing

1.35

 getting

1.35

 being

1.34

Activations Density 0.523%

phrases related to comparison or contrast

phrases that indicate assessments or evaluations of situations

No Comments

No Known Activations