INDEX

Explanations

descriptions of something being completely filled or packed with something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "filled."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

rf

-0.77

peer

-0.76

framework

-0.74

nature

-0.72

hawks

-0.71

Friends

-0.66

cdn

-0.66

 Surveillance

-0.65

arta

-0.65

 messenger

-0.64

POSITIVE LOGITS

 fills

0.95

 vacancies

0.80

owship

0.79

 filled

0.78

fill

0.76

ragon

0.76

 fill

0.75

 shortages

0.75

ummer

0.74

 filling

0.74

Activations Density 0.011%

descriptions of something being completely filled or packed with something

instances of the word "filled."

No Comments

No Known Activations