INDEX

Explanations

phrases related to walking or movement along with physical objects or locations

oai_token-act-pair · gpt-3.5-turbo

references to walking and locomotion, especially in relation to robots and physical movement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ATF

-0.94

 Auction

-0.82

 Firefly

-0.80

 Saban

-0.78

ynski

-0.77

 Bannon

-0.77

 Avalon

-0.75

 Whedon

-0.74

 Sauce

-0.74

 Tarant

-0.73

POSITIVE LOGITS

 walking

1.86

jog

1.70

 walk

1.69

walking

1.68

 Walking

1.63

 stride

1.59

Walk

1.54

 Walk

1.53

 stroll

1.52

 bicy

1.47

Activations Density 0.374%

phrases related to walking or movement along with physical objects or locations

references to walking and locomotion, especially in relation to robots and physical movement

No Comments

No Known Activations