INDEX

Explanations

terms related to unresolved or unanswered issues and mysteries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

uv

-0.81

upiter

-0.79

osterone

-0.77

adder

-0.73

udo

-0.72

bourg

-0.72

acus

-0.71

onica

-0.71

opt

-0.71

adier

-0.70

POSITIVE LOGITS

 questions

0.93

 unanswered

0.84

 Questions

0.82

 captcha

0.76

 mysteries

0.75

Answer

0.74

 question

0.71

 Mysteries

0.70

 pregnancies

0.69

 unexplained

0.67

Activations Density 0.013%

terms related to unresolved or unanswered issues and mysteries

No Comments

No Known Activations