INDEX

Explanations

phrases or names containing the word "rit"

oai_token-act-pair · gpt-3.5-turbo

references to individuals and their works

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

keye

-0.70

emetery

-0.68

nect

-0.66

ĨĴ

-0.64

ARK

-0.62

¶ħ

-0.60

ŀ

-0.60

Err

-0.58

©¶æ

-0.58

 Nationwide

-0.56

POSITIVE LOGITS

sch

0.94

sis

0.94

chard

0.93

ude

0.85

ual

0.83

0.80

tered

0.79

UAL

0.79

grain

0.78

ter

0.76

Activations Density 0.037%

phrases or names containing the word "rit"

references to individuals and their works

No Comments

No Known Activations