INDEX

Explanations

phrases related to product launches or releases

oai_token-act-pair · gpt-3.5-turbo

references to product launches

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

felt

-0.76

meyer

-0.74

Notes

-0.72

ergic

-0.70

 ILCS

-0.70

oen

-0.70

SPONSORED

-0.69

washed

-0.67

respect

-0.66

 stitches

-0.66

POSITIVE LOGITS

 launch

3.68

 Launch

2.64

 launches

2.60

launch

2.51

 launching

2.39

Launch

2.29

 launched

1.93

Laun

1.88

 rollout

1.82

laun

1.70

Activations Density 0.016%

phrases related to product launches or releases

references to product launches

No Comments

No Known Activations