INDEX

Explanations

words and phrases related to logic and reasoning

oai_token-act-pair · gpt-3.5-turbo

references to logic and its applications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Volunte

-0.74

avez

-0.71

ometown

-0.69

 Sitting

-0.66

emale

-0.65

lain

-0.65

 Leopard

-0.63

hold

-0.63

Shar

-0.62

 national

-0.62

POSITIVE LOGITS

 logic

1.03

matical

0.89

 underpin

0.82

matic

0.81

ophical

0.79

ical

0.78

istical

0.78

 Logic

0.78

istically

0.77

 reasoning

0.76

Activations Density 0.019%

words and phrases related to logic and reasoning

references to logic and its applications

No Comments

No Known Activations