INDEX

Explanations

law enforcement-related events and names, particularly related to sheriff's offices 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

furt

-0.92

acles

-0.90

ipedia

-0.89

isks

-0.82

warts

-0.81

aeus

-0.79

htar

-0.78

iable

-0.77

eches

-0.76

ihad

-0.74

POSITIVE LOGITS

 Sheriff

1.36

 sheriff

1.29

 commissioners

1.09

 Jail

1.07

 Commissioners

1.05

 Clerk

1.04

 Supervisor

1.03

 coroner

1.02

 courthouse

0.99

 Superior

0.99

Activations Density 0.647%

law enforcement-related events and names, particularly related to sheriff's offices and criminal activities

No Comments

No Known Activations