INDEX

Explanations

numerical measurements and technical abbreviations

oai_token-act-pair · gpt-3.5-turbo

patterns related to numerical data and statistics

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

 eleph

-0.78

 citiz

-0.78

hement

-0.69

 residences

-0.68

 occas

-0.64

 rooting

-0.64

 tremend

-0.63

 footing

-0.61

 flats

-0.61

 newcom

-0.61

POSITIVE LOGITS

↵

0.88

Python

0.86

Merit

0.86

Í

0.80

ETF

0.78

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.75

posted

0.72

ÂłÂłÂł

0.72

Thu

0.71

Ì¶

0.71

Activations Density 0.289%

numerical measurements and technical abbreviations

patterns related to numerical data and statistics

No Comments

No Known Activations