INDEX

Explanations

terms preceded by the word "also"

oai_token-act-pair · gpt-3.5-turbo

the word "also" to indicate additional information or items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

jam

-0.71

pering

-0.69

 Guru

-0.66

Aren

-0.65

 Sunset

-0.64

assic

-0.63

Mom

-0.62

agus

-0.62

 Closing

-0.62

 Canaver

-0.62

POSITIVE LOGITS

 optionally

0.88

 confir

0.83

yrinth

0.82

 anecd

0.81

 includ

0.78

 incorporate

0.76

 speculate

0.75

 indirectly

0.74

 contribute

0.72

 include

0.72

Activations Density 0.038%

terms preceded by the word "also"

the word "also" to indicate additional information or items

No Comments

No Known Activations