INDEX

Explanations

instances of words related to deception or fake representation

oai_token-act-pair · gpt-3.5-turbo

terms related to deception and the processes of certification or verification

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

 Avalon

-0.81

 Oaks

-0.73

 Aber

-0.72

 Madness

-0.71

 Kingdoms

-0.68

 Wonderland

-0.68

 Dome

-0.66

 Franch

-0.65

 setback

-0.65

 Innov

-0.65

POSITIVE LOGITS

ified

1.70

ifying

1.69

ifiers

1.57

ification

1.52

ifier

1.50

ifies

1.48

ifiable

1.47

ific

1.43

ifications

1.39

eness

1.34

Activations Density 0.038%

instances of words related to deception or fake representation

terms related to deception and the processes of certification or verification

No Comments

No Known Activations