INDEX

Explanations

references to actions or situations where someone is left, departed, or set free

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.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

ALLY

-0.75

alist

-0.69

 Cosponsors

-0.67

issance

-0.67

andan

-0.66

wcsstore

-0.66

oys

-0.65

onse

-0.63

alez

-0.62

wana

-0.61

POSITIVE LOGITS

overs

1.25

 undone

1.16

 untreated

0.92

 behind

0.92

 unfinished

0.89

 unanswered

0.87

 Behind

0.86

wich

0.85

 unexpl

0.81

 intact

0.80

Activations Density 0.462%

references to actions or situations where someone is left, departed, or set free

No Comments

No Known Activations