INDEX

Explanations

phrases related to political and social discussions

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Nav

-0.64

 Siber

-0.62

ãĤª

-0.60

translation

-0.56

 Dover

-0.56

 Airbus

-0.54

enge

-0.54

Lv

-0.53

ILA

-0.53

Load

-0.53

POSITIVE LOGITS

 fame

0.89

 deserve

0.81

 were

0.80

 testified

0.78

 testify

0.78

 participated

0.77

 rejoice

0.76

 succumbed

0.76

 tended

0.76

 perished

0.75

Activations Density 3.827%

phrases related to political and social discussions

No Comments

No Known Activations