INDEX

Explanations

numerical values and quantities

oai_token-act-pair · gpt-3.5-turbo

references to totality or completeness in a context

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

EStream

-0.84

ammy

-0.73

HAHAHAHA

-0.72

 CrossRef

-0.70

CHA

-0.69

universal

-0.68

bringing

-0.68

reverse

-0.68

multipl

-0.66

ipolar

-0.66

POSITIVE LOGITS

 components

1.24

 belongings

1.20

 pieces

1.19

 items

1.17

 occupants

1.17

 passengers

1.14

 equipment

1.13

 details

1.13

 materials

1.12

 guests

1.11

Activations Density 0.291%

numerical values and quantities

references to totality or completeness in a context

No Comments

No Known Activations