INDEX

Explanations

references to personal or hidden identity

oai_token-act-pair · gpt-3.5-turbo

mentions of identity, particularly in contexts involving revelation or concealment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

oys

-0.68

Vaj

-0.67

orld

-0.64

 Spoon

-0.64

 Pigs

-0.64

baugh

-0.63

ittle

-0.63

iths

-0.61

âķĲâķĲ

-0.61

ithing

-0.59

POSITIVE LOGITS

 identity

1.01

 Identity

0.97

 identities

0.96

entity

0.89

 è£ıè

0.88

 Theft

0.85

iosyncr

0.82

ativity

0.74

 credentials

0.74

ãĥĹ

0.74

Activations Density 0.022%

references to personal or hidden identity

mentions of identity, particularly in contexts involving revelation or concealment

No Comments

No Known Activations