INDEX

Explanations

terms related to primary items, conditions, or areas of focus

oai_token-act-pair · gpt-3.5-turbo

references to primary concepts or entities within various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

INESS

-0.77

EEK

-0.75

estern

-0.71

 Tycoon

-0.70

OTO

-0.70

 Doodle

-0.65

ARDS

-0.64

UGE

-0.64

 Braun

-0.63

orse

-0.62

POSITIVE LOGITS

 careg

0.82

tenance

0.80

ities

0.79

 distingu

0.78

stay

0.77

 antagonist

0.75

ignment

0.75

 primary

0.74

 objective

0.71

 distinguishing

0.71

Activations Density 0.013%

terms related to primary items, conditions, or areas of focus

references to primary concepts or entities within various contexts

No Comments

No Known Activations