INDEX

Explanations

names of individuals associated with police incidents and violence

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

ahime

-0.75

angular

-0.70

Xi

-0.68

 Pixie

-0.67

FU

-0.66

 Agric

-0.66

 spons

-0.65

lishes

-0.64

 Dragons

-0.64

Kings

-0.64

POSITIVE LOGITS

 murdered

1.02

 gunned

0.95

 victims

0.92

 massac

0.89

 murdering

0.85

 murderer

0.85

 murder

0.84

 unarmed

0.83

Jr

0.83

 killed

0.82

Activations Density 0.053%

names of individuals associated with police incidents and violence

No Comments

No Known Activations