INDEX

Explanations

mentions of someone being given permission or approval for something by an authority figure

oai_token-act-pair · gpt-3.5-turbo

instances of the word "granted."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

psc

-0.77

ping

-0.75

enegger

-0.72

sis

-0.71

 Sins

-0.71

ateg

-0.70

ategy

-0.69

borgh

-0.68

tes

-0.67

econom

-0.66

POSITIVE LOGITS

 amnesty

1.01

 immunity

0.97

 forgiveness

0.81

 granted

0.80

 permission

0.80

 asylum

0.79

 access

0.76

 renewal

0.76

 acceptance

0.75

uthor

0.75

Activations Density 0.022%

mentions of someone being given permission or approval for something by an authority figure

instances of the word "granted."

No Comments

No Known Activations