INDEX

Explanations

instances of the word "functional"

oai_token-act-pair · gpt-3.5-turbo

references to functionality and efficiency in various contexts

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

 Steele

-0.78

 accuser

-0.76

udge

-0.71

 Tyson

-0.70

Sea

-0.69

raise

-0.68

storm

-0.66

Wen

-0.65

 temp

-0.64

 Gomez

-0.64

POSITIVE LOGITS

 functional

3.41

functional

2.91

 Functional

2.51

 functionally

1.79

 functioning

1.74

Function

1.59

function

1.44

 functionality

1.43

 Function

1.38

 function

1.38

Activations Density 0.014%

instances of the word "functional"

references to functionality and efficiency in various contexts

No Comments

No Known Activations