INDEX

Explanations

the word "punish" or variations of it, indicating a focus on punishment or punitive actions

oai_token-act-pair · gpt-3.5-turbo

terms related to punishment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gow

-0.86

Champ

-0.81

coe

-0.73

sky

-0.71

eds

-0.70

aug

-0.68

clus

-0.68

pres

-0.68

park

-0.68

 Herz

-0.67

POSITIVE LOGITS

 punished

1.13

 punish

1.11

 harshly

1.08

 punishments

0.95

 punishing

0.95

 punishment

0.94

 unfairly

0.82

 sanction

0.81

 punitive

0.79

 penalties

0.79

Activations Density 0.018%

the word "punish" or variations of it, indicating a focus on punishment or punitive actions

terms related to punishment

No Comments

No Known Activations