INDEX

Explanations

words related to the concept of stopping or ending

oai_token-act-pair · gpt-3.5-turbo

references to the term "cease" in various contexts, particularly regarding cessation or interruptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

OTE

-0.77

ramid

-0.75

 Inquisitor

-0.74

é¾įå

-0.71

raviolet

-0.71

rium

-0.71

Lab

-0.69

 Aval

-0.68

uries

-0.67

Choice

-0.67

POSITIVE LOGITS

 cease

1.26

 ceases

0.95

rences

0.91

 seiz

0.85

fires

0.83

 cessation

0.82

ĸļ

0.80

 ceased

0.80

 hostilities

0.78

EngineDebug

0.77

Activations Density 0.005%

words related to the concept of stopping or ending

references to the term "cease" in various contexts, particularly regarding cessation or interruptions

No Comments

No Known Activations