INDEX

Explanations

references to specific actions or events that may be controversial or involve some serious implications

oai_token-act-pair · gpt-3.5-turbo

mentions of accountability and blame regarding actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

lishes

-0.74

alg

-0.71

gae

-0.70

 perenn

-0.69

Zone

-0.67

eki

-0.65

erenn

-0.63

chall

-0.62

igor

-0.62

iasis

-0.60

POSITIVE LOGITS

 wrongdoing

1.09

 intentionally

0.93

 anything

0.93

 improper

0.92

 malice

0.89

 inappropriately

0.88

 intentional

0.88

 deliberately

0.87

 unlawfully

0.86

 beforehand

0.86

Activations Density 0.430%

references to specific actions or events that may be controversial or involve some serious implications

mentions of accountability and blame regarding actions or events

No Comments

No Known Activations