INDEX

Explanations

terms related to bullying or being bullied

oai_token-act-pair · gpt-3.5-turbo

terms related to bullying and its effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Donation

-0.77

aeda

-0.76

ittance

-0.74

arbon

-0.72

ossier

-0.72

vid

-0.70

ixed

-0.70

cession

-0.70

ixture

-0.69

ournal

-0.69

POSITIVE LOGITS

 bullies

1.24

 bullying

1.19

 bully

1.08

 bullied

1.08

 behav

0.85

ãħĭ

0.82

pul

0.80

kids

0.77

 slurs

0.71

HAEL

0.69

Activations Density 0.012%

terms related to bullying or being bullied

terms related to bullying and its effects

No Comments

No Known Activations