INDEX

Explanations

first-person pronouns combined with verbs indicating mental processing

oai_token-act-pair · gpt-3.5-turbo

references to personal experiences and relationships

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

advertisement

-0.77

 Coalition

-0.72

 Siege

-0.69

 reperc

-0.64

 âĩ

-0.63

 omission

-0.62

 Pastebin

-0.60

itely

-0.59

 Advertising

-0.58

 Directors

-0.58

POSITIVE LOGITS

atic

0.83

erk

0.76

ansk

0.76

hers

0.73

atically

0.73

atche

0.69

self

0.69

eps

0.68

Osw

0.68

selves

0.67

Activations Density 0.179%

first-person pronouns combined with verbs indicating mental processing

references to personal experiences and relationships

No Comments

No Known Activations