INDEX

Explanations

statements where someone knows something

oai_token-act-pair · gpt-3.5-turbo

references to knowledge and awareness

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

 pending

-0.64

 dramatic

-0.60

 critical

-0.59

 negative

-0.59

 additional

-0.59

 desk

-0.56

 incorporation

-0.56

due

-0.55

 optional

-0.55

coming

-0.55

POSITIVE LOGITS

 knows

3.34

 understands

2.20

 knew

2.14

 remembers

1.81

 know

1.77

 learns

1.76

 realizes

1.73

know

1.66

 thinks

1.64

 KNOW

1.63

Activations Density 0.016%

statements where someone knows something

references to knowledge and awareness

No Comments

No Known Activations