INDEX

Explanations

proper nouns or names associated with notable individuals and events

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

uyomi

-0.76

vana

-0.74

ypes

-0.63

 Leilan

-0.61

 separat

-0.58

ratom

-0.58

maxwell

-0.57

 inhibitor

-0.56

 vulner

-0.56

 inhibitors

-0.55

POSITIVE LOGITS

baugh

0.91

entin

0.83

uez

0.76

zl

0.76

eve

0.68

 Santos

0.66

port

0.64

ford

0.64

dal

0.61

ez

0.60

Activations Density 0.016%

proper nouns or names associated with notable individuals and events

No Comments

No Known Activations