INDEX

Explanations

positions or directions related to someone's location in an image or scene

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the position of individuals in a group or scene

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

humans

-0.71

morrow

-0.66

izens

-0.62

iom

-0.62

."

-0.61

©¶æ¥µ

-0.59

Recommend

-0.59

Rule

-0.58

universal

-0.58

POSITIVE LOGITS

 foreground

0.95

 flanked

0.94

 pictured

0.86

 photo

0.85

 Photo

0.83

 poses

0.82

 caption

0.81

 pose

0.81

 PHOTO

0.79

 screenshot

0.77

Activations Density 0.052%

positions or directions related to someone's location in an image or scene

phrases indicating the position of individuals in a group or scene

No Comments

No Known Activations