INDEX

Explanations

references to time or chronology

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

nington

-0.67

afety

-0.67

``

-0.65

amina

-0.64

terness

-0.64

angan

-0.63

ococ

-0.63

oke

-0.61

=(

-0.60

ynchron

-0.60

POSITIVE LOGITS

 eras

1.00

 iterations

0.96

 iteration

0.92

era

0.87

 editions

0.86

 predecessors

0.82

 incarnation

0.81

 edition

0.77

 epoch

0.76

iator

0.74

Activations Density 0.078%

references to time or chronology

No Comments

No Known Activations