INDEX

Explanations

mathematical terms related to probability distributions and functions

oai_token-act-pair · gpt-3.5-turbo

specific mathematical functions and notations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Jenner

-0.78

gur

-0.76

xit

-0.73

 Fargo

-0.69

tsky

-0.67

 Kardashian

-0.67

 indictment

-0.65

 Cycling

-0.65

 Memories

-0.64

 Borders

-0.63

POSITIVE LOGITS

ĩ

1.21

Ĩ

1.21

ģ

1.20

Ģ

1.20

ĥ

1.18

ī

1.04

Ĭ

1.02

---------------

0.94

opher

0.93

Ĥ

0.92

Activations Density 0.011%

mathematical terms related to probability distributions and functions

specific mathematical functions and notations

No Comments

No Known Activations