INDEX

Explanations

words related to choices or preferences

oai_token-act-pair · gpt-3.5-turbo

discussions around the interaction of various systems and the implications of choices within societal and governmental structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

estern

-0.67

oso

-0.64

NOT

-0.61

ãĥ´

-0.60

 suspic

-0.58

NOT

-0.58

 thous

-0.57

repre

-0.56

 notwithstanding

-0.56

 INCLUD

-0.53

POSITIVE LOGITS

 itself

1.10

 altogether

1.05

 anymore

0.98

 outright

0.91

 themselves

0.86

nor

0.83

 oneself

0.81

 necessarily

0.78

 himself

0.77

 ourselves

0.75

Activations Density 0.554%

words related to choices or preferences

discussions around the interaction of various systems and the implications of choices within societal and governmental structures

No Comments

No Known Activations