INDEX

Explanations

references to tragic events or individuals

oai_token-act-pair · gpt-3.5-turbo

references to victims and historical narratives surrounding them

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

 Leilan

-0.64

flows

-0.63

 constraints

-0.62

 stamp

-0.60

 Advent

-0.59

 Panther

-0.58

 Hath

-0.57

 waivers

-0.57

backer

-0.56

PRE

-0.56

POSITIVE LOGITS

orian

2.03

ory

1.86

orians

1.85

orical

1.80

oria

1.67

orious

1.67

orically

1.64

oire

1.63

ories

1.63

oric

1.63

Activations Density 0.067%

references to tragic events or individuals

references to victims and historical narratives surrounding them

No Comments

No Known Activations