INDEX

Explanations

settings or features related to different modes of operation within a system

oai_token-act-pair · gpt-3.5-turbo

terms related to modes or settings in a technical or gameplay context

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

owan

-0.77

mony

-0.75

roma

-0.72

 Aires

-0.71

enegger

-0.71

bones

-0.70

riad

-0.68

ä½ľ

-0.68

enda

-0.68

mong

-0.65

POSITIVE LOGITS

 Shift

0.76

 mode

0.73

quit

0.73

etting

0.71

 upgr

0.70

 enabled

0.68

 Mode

0.65

 Mayhem

0.63

 reversal

0.63

 selector

0.63

Activations Density 0.017%

settings or features related to different modes of operation within a system

terms related to modes or settings in a technical or gameplay context

No Comments

No Known Activations