INDEX

Explanations

the term "perpetrators."

oai_token-act-pair · gpt-3.5-turbo

terms related to individuals responsible for crimes or wrongdoing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Trade

-0.66

ete

-0.65

Medic

-0.62

 retirement

-0.60

 interesting

-0.59

 Condition

-0.59

 Hover

-0.58

 seaw

-0.58

 Premium

-0.58

 Rounds

-0.58

POSITIVE LOGITS

 perpetrators

3.31

 perpetrator

2.98

 culprit

2.65

 abusers

2.07

 rapists

2.06

 abuser

1.99

 assailants

1.95

 killers

1.88

 offenders

1.84

 murderers

1.81

Activations Density 0.029%

the term "perpetrators."

terms related to individuals responsible for crimes or wrongdoing

No Comments

No Known Activations