INDEX

Explanations

Descriptions related to various events or situations happening

oai_token-act-pair · gpt-3.5-turbo

themes related to deception and manipulation in narrative contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Â´

-0.66

He

-0.65

isoft

-0.64

buquerque

-0.64

âľ

-0.63

bool

-0.63

HCR

-0.61

His

-0.60

iership

-0.60

VPN

-0.60

POSITIVE LOGITS

 themselves

1.30

 their

1.24

 theirs

1.03

their

1.01

 THEIR

0.90

 they

0.85

 Their

0.84

Their

0.77

selves

0.76

 apiece

0.69

Activations Density 0.900%

Descriptions related to various events or situations happening

themes related to deception and manipulation in narrative contexts

No Comments

No Known Activations