INDEX

Explanations

phrases indicating the source or origin of information or perspective

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "from."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ratulations

-0.84

hai

-0.72

merce

-0.71

irm

-0.71

isode

-0.69

quote

-0.67

 mathemat

-0.66

aca

-0.66

BSD

-0.63

lied

-0.62

POSITIVE LOGITS

 afar

1.55

 whence

1.18

 elsewhere

0.92

 scratch

0.91

 inside

0.90

 anywhere

0.89

 thence

0.88

 somewhere

0.85

 everywhere

0.84

 abroad

0.83

Activations Density 0.168%

phrases indicating the source or origin of information or perspective

occurrences of the word "from."

No Comments

No Known Activations