INDEX

Explanations

discussion around social and political issues, particularly related to feminism, the internet, entitlement, and societal expectations

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

 suicides

-0.59

neutral

-0.58

aldehyde

-0.56

 horm

-0.56

ars

-0.55

 attrition

-0.55

 confirmation

-0.54

 notably

-0.53

wash

-0.53

isol

-0.53

POSITIVE LOGITS

 entails

0.78

 entail

0.76

ACA

0.72

aca

0.72

 Anyway

0.71

Cry

0.70

SourceFile

0.66

FFER

0.66

FACE

0.65

 versus

0.65

Activations Density 1.964%

discussion around social and political issues, particularly related to feminism, the internet, entitlement, and societal expectations

No Comments

No Known Activations