INDEX

Explanations

mentions of arrests, charges, and criminal activities

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

issance

-0.76

 Remastered

-0.74

ankind

-0.71

equal

-0.69

emic

-0.67

split

-0.65

ertain

-0.65

dust

-0.65

ctory

-0.63

 spons

-0.61

POSITIVE LOGITS

 Warrant

0.91

llah

0.81

 felon

0.80

 arrest

0.80

 handcuffed

0.80

 arrested

0.78

 Officers

0.78

 apprehended

0.77

 warrants

0.77

 warrant

0.73

Activations Density 0.643%

mentions of arrests, charges, and criminal activities

No Comments

No Known Activations