INDEX

Explanations

phrases related to news headlines or reports of criminal activities

oai_token-act-pair · gpt-3.5-turbo

references to crime and related incidents

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iencies

-0.77

issance

-0.77

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.71

 Tolkien

-0.70

 stagnation

-0.70

ointment

-0.69

 ageing

-0.68

Frameworks

-0.66

 immortality

-0.66

kefeller

-0.65

POSITIVE LOGITS

TMZ

1.28

 eyewitness

1.24

 cops

1.12

Police

1.12

 NYPD

1.08

 Inquis

1.07

CTV

1.07

VIDEO

1.06

 VIDEO

1.05

 Police

1.04

Activations Density 0.699%

phrases related to news headlines or reports of criminal activities

references to crime and related incidents

No Comments

No Known Activations