INDEX

Explanations

references to self-identification or self-references in the text

oai_token-act-pair · gpt-3.5-turbo

references to self-identification

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

pour

-0.76

onga

-0.75

ibaba

-0.71

icion

-0.70

rought

-0.68

edia

-0.68

iard

-0.67

iens

-0.66

oos

-0.66

heny

-0.63

POSITIVE LOGITS

selves

1.06

 worshipped

0.81

 selves

0.77

æ³

0.73

self

0.70

 è£ıè

0.69

ç¥ŀ

0.68

 creatively

0.68

 adherent

0.67

 acknow

0.67

Activations Density 0.035%

references to self-identification or self-references in the text

references to self-identification

No Comments

No Known Activations