INDEX

Explanations

instances of people or entities joining something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "joined."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ester

-0.69

abytes

-0.65

otropic

-0.63

 calculation

-0.61

information

-0.60

astrous

-0.60

relative

-0.60

ager

-0.59

 embryos

-0.59

abyte

-0.59

POSITIVE LOGITS

 joined

0.89

ĸļ

0.85

Join

0.84

 joins

0.79

join

0.79

 Join

0.78

ç·

0.77

 join

0.77

 joining

0.73

ertodd

0.71

Activations Density 0.009%

instances of people or entities joining something

instances of the word "joined."

No Comments

No Known Activations