INDEX

Explanations

mentions of and references to penalties, especially related to legal or sporting contexts

oai_token-act-pair · gpt-3.5-turbo

references to penalties, particularly in a legal or regulatory context

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

agos

-0.86

 Niet

-0.73

 Begin

-0.66

idon

-0.64

artz

-0.64

 Explorer

-0.64

 Digest

-0.63

atz

-0.63

Untitled

-0.62

 Generations

-0.62

POSITIVE LOGITS

 penalty

3.96

 penalties

2.98

 Penalty

2.90

 penal

1.75

 punishment

1.74

 punishments

1.53

 fines

1.34

 bonus

1.29

 forfeiture

1.22

onement

1.20

Activations Density 0.005%

mentions of and references to penalties, especially related to legal or sporting contexts

references to penalties, particularly in a legal or regulatory context

No Comments

No Known Activations