INDEX

Explanations

details or information related to specific reports, books, hacks, court cases, removals, services, campaigns, accounts, carnage, connections, references, releases, polls, sitcoms, data centers, projects, Korean ambitions, bill registrations, data centers, and communication facilitation

oai_token-act-pair · gpt-3.5-turbo

mentions of reports, legal matters, or incidents involving investigations and their associated details

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

laughs

-0.82

 Learns

-0.69

`.

-0.67

 Kills

-0.61

Laughs

-0.60

 Gets

-0.60

$.

-0.60

''.

-0.60

];

-0.59

!".

-0.57

POSITIVE LOGITS

are

1.41

 vary

1.36

 abound

1.35

 were

1.32

 aren

1.27

 include

1.24

 weren

1.24

 differ

1.24

 indicate

1.16

 differed

1.16

Activations Density 0.750%

details or information related to specific reports, books, hacks, court cases, removals, services, campaigns, accounts, carnage, connections, references, releases, polls, sitcoms, data centers, projects, Korean ambitions, bill registrations, data centers, and communication facilitation

mentions of reports, legal matters, or incidents involving investigations and their associated details

No Comments

No Known Activations