INDEX

Explanations

proper nouns related to a specific individual or organization

oai_token-act-pair · gpt-3.5-turbo

references to specific political figures and entities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mem

-0.71

wolf

-0.69

 wolves

-0.65

fort

-0.65

 macros

-0.65

 Fantasy

-0.65

 wolf

-0.63

ops

-0.63

 trapped

-0.62

 Ghost

-0.62

POSITIVE LOGITS

 Bhar

4.50

 Gujar

1.54

 Maharashtra

1.43

Aad

1.42

 Vish

1.38

 Brah

1.35

 Gujarat

1.33

Bh

1.33

 Laksh

1.33

 Gupta

1.30

Activations Density 0.026%

proper nouns related to a specific individual or organization

references to specific political figures and entities

No Comments

No Known Activations