INDEX

Explanations

expressions of positivity and delight

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

 objectionable

-0.78

acist

-0.77

ById

-0.71

 overriding

-0.70

 wills

-0.68

uyomi

-0.63

lest

-0.63

untled

-0.62

 tant

-0.62

 leve

-0.61

POSITIVE LOGITS

clus

0.83

 hear

0.72

 Featured

0.70

 congr

0.70

 seeing

0.69

someone

0.68

Featured

0.66

days

0.66

DN

0.63

noon

0.62

Activations Density 1.347%

expressions of positivity and delight

No Comments

No Known Activations