INDEX

Explanations

phrases related to adding or appending items

oai_token-act-pair · gpt-3.5-turbo

references to the term "Add" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 outnumbered

-0.69

 firsthand

-0.62

 resemb

-0.61

âĢķ

-0.59

 fair

-0.59

é¾įå¥ĳå£«

-0.59

PATH

-0.57

 similarities

-0.55

Mub

-0.55

 bounce

-0.55

POSITIVE LOGITS

endum

1.60

ictive

1.55

icts

1.55

ressing

1.53

ictions

1.52

icted

1.51

resses

1.49

ition

1.44

iction

1.44

ict

1.39

Activations Density 0.027%

phrases related to adding or appending items

references to the term "Add" in various contexts

No Comments

No Known Activations