INDEX

Explanations

phrases related to projections and estimates of future trends or values

oai_token-act-pair · gpt-3.5-turbo

phrases related to projections and expectations regarding data trends

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

soType

-0.69

sonian

-0.63

 recalls

-0.63

 listens

-0.62

 Takeru

-0.62

 reflections

-0.61

Syd

-0.60

 Tess

-0.60

 reminders

-0.60

 demos

-0.59

POSITIVE LOGITS

 exceed

1.39

 increase

1.23

 soar

1.21

 accelerate

1.19

 surpass

1.17

 decrease

1.14

 overtake

1.13

 reduce

1.11

 generate

1.10

 comprise

1.10

Activations Density 0.158%

phrases related to projections and estimates of future trends or values

phrases related to projections and expectations regarding data trends

No Comments

No Known Activations