INDEX

Explanations

phrases related to supervision and monitoring

oai_token-act-pair · gpt-3.5-turbo

terms related to supervision and oversight

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tr

-0.70

grave

-0.69

ted

-0.69

false

-0.68

anne

-0.68

fixes

-0.67

UES

-0.67

utter

-0.67

tons

-0.66

tery

-0.66

POSITIVE LOGITS

 supervision

1.21

 supervised

1.04

 superv

0.97

 Instruct

0.84

 oversee

0.83

 probation

0.83

vised

0.73

 Instructor

0.72

 overseen

0.71

 confir

0.71

Activations Density 0.008%

phrases related to supervision and monitoring

terms related to supervision and oversight

No Comments

No Known Activations