INDEX

Explanations

terms related to negative events or controversies, specifically around political or social issues

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

abase

-0.66

UL

-0.58

 strengths

-0.56

enh

-0.56

 Atkinson

-0.55

amental

-0.55

iser

-0.54

 Origin

-0.54

 Edmund

-0.53

uay

-0.53

POSITIVE LOGITS

 suddenly

0.86

 somehow

0.85

 bothered

0.73

ammed

0.72

so

0.70

 openly

0.69

agos

0.69

 juxtap

0.66

 singled

0.65

 resorted

0.65

Activations Density 0.400%

terms related to negative events or controversies, specifically around political or social issues

No Comments

No Known Activations