INDEX

Explanations

words related to moral judgment and criticism towards authority figures

oai_token-act-pair · gpt-3.5-turbo

references to political accountability and legal proceedings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 seamless

-0.73

 versatility

-0.72

 srfAttach

-0.69

 transitions

-0.68

 seamlessly

-0.68

 Transition

-0.67

 combination

-0.66

 effortlessly

-0.64

 Dynamo

-0.63

izontal

-0.62

POSITIVE LOGITS

 blaming

1.33

 condemning

1.30

 accusing

1.26

 criticizing

1.23

 complain

1.20

 criticize

1.19

 condemn

1.18

 complaining

1.18

 accuse

1.15

 blame

1.15

Activations Density 0.533%

words related to moral judgment and criticism towards authority figures

references to political accountability and legal proceedings

No Comments

No Known Activations