INDEX

Explanations

references to criminal activities and legal proceedings surrounding a crime

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

leaf

-0.68

ooks

-0.64

ove

-0.64

 princip

-0.64

 authoritarian

-0.63

efficiency

-0.62

raltar

-0.60

 taxing

-0.60

binding

-0.59

andel

-0.59

POSITIVE LOGITS

istics

0.88

izes

0.88

izer

0.85

izers

0.82

vict

0.81

ization

0.78

 Victim

0.77

 inflicted

0.77

 victim

0.77

hood

0.76

Activations Density 0.030%

references to criminal activities and legal proceedings surrounding a crime

No Comments

No Known Activations