INDEX

Explanations

Words related to technical specifications, such as acronyms and codes

oai_token-act-pair · gpt-3.5-turbo

occurrences of the abbreviation "GR" and its variations, indicating a focus on a specific topic or classification system

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

lihood

-0.82

ters

-0.72

than

-0.71

hered

-0.71

tern

-0.70

que

-0.70

\":

-0.69

manship

-0.69

 Learns

-0.66

 Wand

-0.66

POSITIVE LOGITS

APH

1.20

ADE

1.03

izont

0.97

ASS

0.93

ANT

0.91

ATOR

0.91

ANCE

0.90

OUND

0.89

OSS

0.88

APE

0.87

Activations Density 0.015%

Words related to technical specifications, such as acronyms and codes

occurrences of the abbreviation "GR" and its variations, indicating a focus on a specific topic or classification system

No Comments

No Known Activations