INDEX

Explanations

phrases indicating something valuable or deserving attention

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

thumbnails

-0.71

gdala

-0.64

Ern

-0.63

 Hilbert

-0.62

 Boards

-0.60

rete

-0.60

oslov

-0.59

 Townsend

-0.59

 Graphics

-0.58

Kat

-0.58

POSITIVE LOGITS

 mentioning

0.89

iness

0.88

folios

0.86

 consideration

0.86

 noting

0.81

keeping

0.79

ily

0.78

 careful

0.75

ighed

0.74

gotten

0.74

Activations Density 0.668%

phrases indicating something valuable or deserving attention

No Comments

No Known Activations