INDEX

Explanations

words related to prisons and incarceration

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

£ı

-0.70

ãĥ£

-0.70

thora

-0.70

yip

-0.68

endor

-0.66

OPA

-0.65

rians

-0.64

soDeliveryDate

-0.63

cs

-0.62

sonian

-0.61

POSITIVE LOGITS

 inmates

1.20

 inmate

1.11

 confinement

1.06

 prisoners

1.02

prison

1.01

 sentences

0.96

 prison

0.95

 jails

0.93

 prisons

0.92

 convict

0.90

Activations Density 0.843%

words related to prisons and incarceration

No Comments

No Known Activations