INDEX

Explanations

phrases related to flags

oai_token-act-pair · gpt-3.5-turbo

references to flags, particularly in a metaphorical or literal context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ther

-0.81

vana

-0.78

aday

-0.78

liv

-0.76

Sco

-0.75

essen

-0.74

mort

-0.71

okin

-0.68

sein

-0.68

 Neuroscience

-0.67

POSITIVE LOGITS

 flag

3.89

flag

2.83

 Flag

2.76

 flags

2.76

Flag

2.48

 Flags

1.96

flags

1.85

Flags

1.78

 flagged

1.58

 banner

1.54

Activations Density 0.010%

phrases related to flags

references to flags, particularly in a metaphorical or literal context

No Comments

No Known Activations