INDEX

Explanations

the acronym "MIT"

oai_token-act-pair · gpt-3.5-turbo

references to the Massachusetts Institute of Technology (MIT)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ãħĭ

-0.75

naire

-0.72

loading

-0.71

wagen

-0.69

 Arpaio

-0.67

boarding

-0.66

ppings

-0.63

finder

-0.63

usterity

-0.61

 ugly

-0.61

POSITIVE LOGITS

 Sloan

1.07

LEY

0.89

 Libraries

0.89

MIT

0.88

CLA

0.86

RE

0.86

MED

0.86

DEF

0.83

ENA

0.82

heastern

0.81

Activations Density 0.022%

the acronym "MIT"

references to the Massachusetts Institute of Technology (MIT)

No Comments

No Known Activations