INDEX

Explanations

phrases related to prisons or incarceration

oai_token-act-pair · gpt-3.5-turbo

references to prisons and the prison system

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

yip

-0.98

lass

-0.77

thora

-0.70

rians

-0.70

ï¸ı

-0.70

omatic

-0.69

rian

-0.69

udden

-0.69

witz

-0.66

zinski

-0.65

POSITIVE LOGITS

 inmates

0.95

 prisons

0.85

 inmate

0.85

 sentences

0.84

 confinement

0.79

prison

0.78

 barr

0.78

 prisoners

0.77

 prison

0.77

house

0.76

Activations Density 0.030%

phrases related to prisons or incarceration

references to prisons and the prison system

No Comments

No Known Activations